Archives for category: Teaching

You, my colleagues, are world-renowned researchers, entrepreneurs, educators, and leaders.   I’d feel presumptuous offering you my thoughts on how to pursue research, entrepreneurship, etc.   I might as well share my thoughts on basketball with Coach K.   Thus, I’m going to focus on an issue that I perhaps spend more time thinking about: why I love my job.  And yes, I do love my job – and not just because I can sit outside with a cup of coffee discussing new research ideas with brilliant students and call it “work.”

As professors in a department like Duke ECE, we have an unusual ability to shape our jobs.  We get to choose our research topics, our collaborators, our students, and the classes we teach.   We control our research group structures, and we can adjust how much time we devote to each aspect of our jobs.   We can even control how much time we devote to work and which hours we choose to work.  

Given that there is more that I’d like to do than I have time for doing well, I have tried to shape my job such that my time and energy are spent on the activities that I find most rewarding.

Mentoring Students

If I had to single out the part of my job that is the most fulfilling, it is seeing my former students succeed.  To paraphrase a co-advisor of mine, our most important products as professors are our students.  We have papers we’re particularly proud of, but we end up being most proud of our students.    

I have had the great joy of mentoring many wonderful PhD students and undergraduates.  I enjoy this part of my job and thus try to maximize my time for it.   (I have had a few students who were difficult and not terribly enjoyable to advise, but they were thankfully in the minority.)  I have intentionally kept my group fairly small—4 or 5 PhD students and 1 or 2 undergraduate research assistants—so that I have enough time to devote to each of them.  My meetings with my students are usually the highlights of my week, and I don’t believe I could manage more students and still have the same kind of relationships with them.  I fully realize that there are faculty, including colleagues here in Duke ECE, who can have more students and easily find time for all of them; I’m just not one of them.  Perhaps my group size will increase when my kids are older and want less of my time.  My group size decision is a conscious trade-off: there are certain projects that just can’t be done with a group of this size.


I very much enjoy teaching students who are genuinely interested in the topic, and I’m happy to devote time for that.  I still get a kick out of that moment when the students realize that computers aren’t magic.   At Duke, I’m extraordinarily lucky to teach a lot of superb students and relatively few slackers.  I have found it tremendously rewarding when former students contact me or visit me to tell me about how they have used what they learned in their jobs and in graduate school.   I do worry about eventually getting bored teaching the same courses over and over, but I hope that alternating among four different courses helps to stave that problem off.

The only part of teaching I dislike—other than grading, which I absolutely hate and thus delegate to TAs—is handling the requests for extensions, special consideration, make-up work, etc.   Early on I discovered I could solve this problem by having a single one-size-fits-all policy (summary: “No”), and I could live with being considered “inflexible” (or other less tactful words one finds in one’s teacher/course evaluations).  Then I discovered that I could pitch this policy as fairness, since how could I possibly judge the relative merits of 50 different excuses in a way that’s fair to all?   Students bought that explanation and it has saved me time that I can devote to other activities.  The only drawback is receiving less fodder for my Fault Tolerant Computing course, because I’m no longer told about as many instances of computers failing (or being stolen or possessed by gremlins) the night before assignments are due.

Service: Committees and Undergraduate Advising

Service is a necessary part of the job, but I doubt there are many of us who joined the department because of a burning desire to serve on the graduate studies committee.   But, given that service is necessary, the key has been finding service jobs where I care about the outcome and where I may have something unique to offer.  I was (relatively) happy to serve on IT committees because I rely on the IT infrastructure here and I’m often among the first ones to break a new service.  I was (relatively) happy to serve on a joint CS/CE committee because the relationship between CS and Computer Engineering (and ECE, as a whole) is important to me.  I was (relatively) happy to serve on faculty search committees because I care about who my new colleagues will be.  All of these service jobs were preferable to alternatives, some of which I wasn’t clever enough to escape.

Early on, I discovered that undergraduate advising was a bimodal experience.  I enjoyed advising ECE/CS double majors, and it was rewarding to be able to offer them useful career guidance and connect them to colleagues in academia and industry.  However, it was miserable to advise BME majors, because I had nothing to offer them.   A typical student question was something like “So, Prof. Sorin, what do you think of BME 273?”  I don’t think about BME 273.   (I offered less grouchy, but no more insightful, responses to the students.)   At one point it occurred to me that I could ask promising ECE/CS double majors in my classes to switch to be my advisees.  The key is making sure I take on enough new ECE/CS advisees so that the DUS doesn’t think I need any other advisees.   I tend to end up with perhaps a few more advisees than I’d otherwise have if I was passive, but the advising is far more rewarding this way.   (Yes, I realize that this approach isn’t scalable, so don’t all start doing this!)


One of the many aspects of Duke that drew me here was seeing that faculty could be highly successful and have families and lives.   I cherish the ability to leave Duke early and be with my kids from around 4:30 until their bedtime, knowing that I can get my work done later.    I consciously categorize my work into “must do at Duke” and “could do from home.”   I like that you, my colleagues, care about what a faculty member accomplishes without needing to compete to see who can spend the most hours visibly working at Duke.

Personal Freedom vs. Departmental Service

So everything is great, right?  We can shape our jobs in a way that we can be successful and apply our limited time and energy towards the activities we find most rewarding.  But what if we all shape our jobs to suit ourselves?  One problem is that we’d be unlikely to have a DGS, given that few of us would prioritize being DGS in terms of how rewarding it is (compared to, say, research).  OK, so we can “incentivize” jobs like this with perks, although I’ve not yet figured out what perks would be required to persuade me to take on that job.  But what about serving on committees?  What about teaching large, core undergraduate courses (instead of small graduate seminars)?  The operation of an academic department like Duke ECE fails Game Theory 101; the incentives are set up in a way that almost completely discourages time spent on service.  We are rewarded for our research contributions—and, to a somewhat lesser extent, our teaching—and thus time spent on service is clearly counter-productive.

So why does our department (usually) function well despite the presence of disincentives to service?  It’s because of you, my conscientious colleagues.  We tend to do “the right thing” even when it’s not in our interests.  It certainly helps that departmental leadership publicly recognizes our service, but I doubt most of us serve on committees because we’ll get recognized for that service.  It seems like we should need to readjust the incentive system to encourage service, but I have not reconciled this goal with the justifiable all-importance of research. 

Concluding Thoughts

Many colleagues of mine at other universities have left academia to take jobs in industry.  Although some of them left for the understandable reason of wanting a new challenge or for financial reasons, others have told me about how they couldn’t find enough time to do the parts of their jobs that they enjoyed the most.  I try to keep that thought in mind when I find myself drifting (“hey, that looks like it could be fun”) instead of being strategic with my time.   Perhaps that’s why I still love my job.


One of the most rewarding aspects of being a professor is guiding students as they identify and achieve their educational, research, and professional goals. This mentorship role is often buried within the faculty mission of research, teaching, and service, but it is a core component of our daily activities and responsibilities. As such, faculty mentorship is one of the strongest resources for creating successful outcomes for graduate students…the challenge is to incorporate such mentorship earlier in the graduate experience.

Socialization is defined as “…the processes through which individuals gain the knowledge, skills, and values necessary for successful entry into a professional career…” [1]. In the traditional Ph.D. program, the faculty research advisor imparts this knowledge to the Ph.D. student, most often in the context of being an independent researcher. Typically, this mentorship occurs once the student has completed all coursework and qualifying exams, and is fully integrated into the research group. Yet, for most Ph.D. students, the foundation for completing the degree and much of their future success is rooted in the mentoring provided during the first two years.

One of the greatest dangers for any graduate student after matriculation, and especially for members of underrepresented groups, is the feeling of isolation or the concern that you do not fit into the culture of your research group, department, and/or chosen field of study. As an African-American woman and member of a group that is sorely underrepresented in science and engineering, I am acutely aware of the perils posed during this critical, initial period of graduate study. From the time of matriculation, Ph.D. students require socialization to their research group, department, and institution; yet, the relationship with the faculty advisor is new and under development, and may present its own set of challenges. Therefore, alternative forms of mentorship are an essential approach to the successful socialization of graduate students.

As I reflect on my experiences as a Ph.D. student at the University of Michigan, I find that in many ways, it was representative of the traditional graduate program structure. I was very fortunate to have had an excellent faculty advisor and mentor; but, this relationship did not fully blossom until well after the qualifying exam in my second year. During the first two years of study, like many other graduate students, my academic life was dominated by coursework, interactions with peers and faculty instructors, and preparation for the qualifying exam. Fortunately, my first two years of graduate study were also characterized by peer mentoring, which provided socialization and a sense of community. Peer mentorship occurred spontaneously because the Applied Physics Program at UofM consisted of a common physics curriculum for all students in the first year and required weekly seminars for the first two years.

These requirements created an atmosphere in which cohorts of Applied Physics students often took courses together and formed study groups, and encouraged first-year and second-year students to interact and share knowledge about navigating the program, selecting research advisors, and passing the qualifying exam. Another important aspect of my experience at UofM was that the student body and faculty in science and engineering included a critical mass of underrepresented minorities that provided role models and successful examples for emulation.

Of course, my graduate school experiences are not unique, and the peer mentoring that was instrumental to my success as a graduate student is not the only model for promoting better socialization through mentorship. The growing recognition of the importance of socialization and mentorship in graduate education is well documented. The 2001 ASHE-ERIC Higher Education Report, “Socialization of Graduate and Professional Students in Higher Education: A Perilous Passage?,” identifies the following trends that necessitate reform of the traditional Ph.D. program structure [1]:
Diversity – One aspect of graduate student socialization is that the “backgrounds and predispositions of prospective graduate students” influence their perceptions of belonging to a potential graduate program and their decisions of whether to attend a given institution or not. The impact of the program structure on the ability to recruit diverse student populations (i.e., underrepresented groups, women & minorities) should be considered.
International Graduate Students – As the graduate student population continues to comprise large percentages of international Ph.D. students, it is important to ensure these students receive professional socialization beyond the routine attention to academics and research.
Professionalism – A successful Ph.D. program should not only prepare students for completion of coursework and the research dissertation, but it should instill in students the qualities and sensibilities of professionalism that they will be expected to display after graduation.
Professionalization – A successful Ph.D. program should also provide some guidance and training on issues that have historically been learned through on-the-job training, such as grant-writing or teaching in the case of academia.
Ethics – While ethics are not always explicitly taught in the context of Ph.D. education, the expectation exists that students will be well-versed in ethics issues during their professional careers.
Technology and Distance Learning – As technology continues to enable distance learning and on-line education, it is important to determine how this will impact socialization and transfer of knowledge, especially with respect to Master’s education.

We can see examples of some of these trends being addressed in our current Ph.D. graduate program, such as the Responsible Conduct of Research (RCR) training required by the graduate school, the technical writing and presentation courses required for international graduate students, and the Ph.D.+ Program offered by the Pratt School of Engineering. Less obvious are the actions we can take as faculty and as a department to better support our Ph.D. students and to promote socialization early in our graduate program. Some reforms have been suggested in higher education literature.

In the same 2001 ASHE-ERIC Higher Education Report, the following suggestions for improvement were made [1]:
Modifying the program – “More collaborative, holistic approaches to learning necessitate systemic change that challenges most existing approaches to graduate and professional study.”
Increasing diversity – “Developing greater flexibility and more options for students so that graduates are more versatile, attracting more women and minority group members, and providing better information about careers continue to be among the major areas of improvement advocated by major national commissions.”
Offering support for students – “Graduate programs will have not only to create more supportive and collaborative environments in the face of increasing diversity but also to sustain them over time.”
Modifying faculty and administrative roles – “The relationship of faculty to students should be interactive, collaborative, open, and mutually evaluative. Relationships need not be power based but should be more interactive with faculty-student, teaching, and research relationships more cooperative.”

Austin provides student recommendations for how graduate programs can provide better socialization, especially in the context of preparing future faculty [2]. The following two suggestions are of particular interest:
More attention to regular mentoring, advising, and feedback – “…faculty members should provide regular, ongoing advising and thorough, periodic feedback and assessment. Assessment should help students determine their progress as scholars and future faculty members. Such advising requires department chairs and graduate deans to work with faculty members to develop effective, mutually respectful, efficient advising relationships. In addition to effective advising, reducing the conflicts between faculty members and graduate students is important.”
Regular and guided reflection – “…graduate students should be encouraged to engage in ongoing, systematic self-reflection. Socialization for doctoral students is largely about making sense of graduate school and the academic career, developing one’s interests and areas of strength, determining how one’s values and commitments relate to those in the profession, and developing one’s own sense of place and competence within that profession. The time and support for reflection are important ingredients in the socialization process.”

Austin synthesizes the student recommendations to propose a modest revision of the typical PhD program [2]:
“A revised doctoral program could begin with an opportunity for entering students to discuss with faculty members their intellectual and professional goals. Though students’ goals often change as they gain experience and learn more about the questions of their fields, the initial assessment could be used to begin focused planning and decision-making. A planning session at time of entry could be followed by annual discussions with a faculty advisor about how the student’s goals are changing and how courses, research, teaching, and other experiences are contributing to progress toward the goals.”

Austin also notes [2]:
“Without a plan these recommendations might appear to add more time to a doctoral program or to the work of already busy faculty members. Yet many of these suggestions involve reorganizing, not adding time.”

Chesler and Chesler describe alternative, gender-informed, mentoring models for women engineering scholars [3]:
“Model 3: Collective Mentoring – Collective mentoring is an evolution of the multiple mentor/single mentee model whereby senior colleagues and the department take responsibility for constructing and maintaining a mentoring team. The entire department or organization must establish and ensure the effective mentoring and performance of graduate students and young professionals. In this way, senior colleagues and the department itself send the message that their progress is a priority concern and may create a departmental climate that overcomes some of the obstacles not only to effective mentoring of women, but also to their effective performance, retention, and advancement.”

Davidson and Foster-Johnson suggest specific actions that departments can take to improve the mentorship of graduate researchers of color [4]. Of particular interest is the suggestion that departments create formal mentoring programs in which faculty are rewarded for participation, which should be voluntary, and that choice should be involved in the matching process with mentors.

It is important to note that one of the positive outcomes of being sensitive to diversity is that the community and departmental culture can improve for everyone when the programmatic structure accounts for vulnerable members of underrepresented groups. As Chesler and Chesler state [3], “Organizational change that creates more egalitarian and caring communities will benefit men as well as women.”

I believe that a stronger sense of community, as well as a stronger faculty presence in and commitment to the development of PhD students during the early years of graduate study, are the best ways to improve departmental culture, to combat attrition, to increase diversity of our graduate student population (specifically students from underrepresented groups), to better serve diverse student populations, and to continue to attract and matriculate the very best PhD students. So, to conclude, I pose the following questions:
1) What structural changes can/should we make to our Ph.D. program to promote peer mentoring and to foster a sense of community?
2) How can we as faculty members be active participants, as mentors and advocates, in the socialization of Ph.D. students after matriculation?

With modest effort, creative thinking, and flexibility, I am certain that we can channel our existing mentorship functions into a force for improvement within our department.

[1] J. C. Weidman, D. J. Twale, and E. L. Stein, “Socialization of Graduate and Professional Students in Higher Education: A Perilous Passage?,” San Francisco, CA 2001.
[2] A. E. Austin, “Preparing the Next Generation of Faculty: Graduate School as Socialization to the Academic Career,” The Journal of Higher Education, vol. 73, pp. 94-122, 2002.
[3] N. C. Chesler and M. A. Chesler, “Gender-informed mentoring strategies for women engineering scholars: On establishing a caring community,” Journal of Engineering Education, vol. 91, 2002.
[4] M. N. Davidson and L. Foster-Johnson, “Mentoring in the Preparation of Graduate Researchers of Color,” Review of Educational Research, vol. 71, pp. 549-574, 2001.

Our students often ask us for recommendations and feedback about their career path and related topics. I find such types of conversations among the most important part of our job as educators. I would not dare to give advice in this letter, after all, as Oscar Wilde said, “I am not young enough to know everything.” Others are certainly much more qualified than me to advise, for example see President Richard H. Brodhead’s recent convocation speech for the class of 2017 ( I simply want to share some of my personal experience and thoughts and invite the interested reader to go for a coffee (or tea in my case) to continue these important and fun conversations. For me academia has been the best job possible, and I will try to convey why.

Let me start by a brief biography since my life experience has without a doubt shaped me into who I am today. I left my country, Uruguay, at the age of 17. Leaving everything behind at such early age is a challenge, one that we as educators must understand since a large population of our students left their homes at a similar age. I first arrived at a Kibbutz in Israel, where I spent 6 marvelous months. Coming from a very socialist background, and after living the majority of my life under a military dictatorship, being in a Kibbutz was not only an incredible way to adapt to a new country, but also a way to experience what I have been reading about and discussing with my youth friends. After that unique experience, I moved to Haifa, and applied to the Technion, Israel Institute of Technology. Like many, I didn’t know what to study; after all, asking 17-18 years old what they want to do when they grow up is in the majority of the cases, a pointless question (I am now 47 and still don’t know what I want to do when I grow up). I knew I loved math (still do), and I thought about applying to the math department. One of my roommates explained to me that it was a mistake, since to get into the math department I had to only get about 75% in the admission exam, and that I should apply to electrical engineering where I needed to get about a perfect score (a simple consequence of supply and demand). So I did, got admitted, and from then, my academic life has been a dream! I spent 4 years working very hard, harder than ever, and got my degree. I loved it so much that I decided to continue to graduate school (or maybe I was running away from the responsibility of a real job?). I was fortunate to be accepted by David Malah to his team, and that was probably my first experience on what a great mentor is and how much he/she can influence the life of a student. Half of what I am today in academia is thanks to David, and the other half to my PhD advisor, Allen Tannenbaum. While I learned a lot about science, engineering and research from them, I learned as much about life. We are still very good friends, and they are the role models for me for the way I want to be with my own students. A good teacher, mentor, and advisor can help you to love what you do.

I loved the Technion, still do; I consider it the best university in the world (sorry Duke, but being second to the Technion is nothing to be ashamed of). I left to MIT for a year, already having a faculty position waiting for me back at the Technion: a dream job. That year was 20 years ago. How I ended up for 3 years at HP Labs in Palo Alto instead of going back to Israel is a long story for another occasion, but this was an industrial experience I recommend to everybody in our field. After that the possibility of moving to the University of Minnesota appeared, and I took it, I don’t regret it at all. I spent 15 great years in Minnesota before coming to Duke in 2012. I had and still have a blast in my job.

I was interviewed a few years ago by the main newspaper in Uruguay on the occasion of receiving the PECASE from the White House (yes, you get to visit the White House, unescorted!), and I was asked why I moved from industry back to academia. I gave many motives, but one (I will give a few more down below) was that I find universities a place where we don’t get old: we keep innovating, meeting new fabulous people, and new and naïve students with great ideas and questions. Students… they can make us so proud!

Let’s continue this line of thought. Every year we teach new students, and hopefully we positively shape their future, like my teachers and mentors helped me to shape mine. Even when we engage in innovative ways of teaching, like MOOCS, we meet new students. While I got engaged with Coursera for many motives, one being the fact that in science and in all new experiments we have to be players and not spectators, without any doubt for me the best part of my Coursera experience were the forums, and to see how naturally students help each other and how we can communicate with people otherwise we would have never had the chance to interact. Knowledge, and in particular the thirst for knowledge, transcends languages, political views, religion, and everything in between.

Academia also permits us to keep innovating scientifically. While all my research in the past 20 years is connected, what I am doing today is very different from what I did 20 years ago, and that is the case for most of my preferred colleagues. I am lucky that I am involved in very challenging research that can be life transformative as well, like our activities in deep brain stimulation (neurosurgery), HIV, and early child behavioral analysis. What can be more fun and rewarding than engaging in challenging activities that have the potential, one step at a time, of helping others?

This brings me to another key reason why I consider this the best job in the world. It allowed me to meet incredible colleagues and human beings. My closer collaborators are also among my best friends, often we first become friends and then collaborators, but sometimes the other way around also happens. I have so much fun, and have learned so much from my collaborators, that is sometimes incredible we get paid to do this (though don’t believe our salaries are so high of course). I don’t want to name all my close friends and collaborators here, we can do that when we get together for that coffee I mentioned above, but let me name one that very sadly, just passed away a few weeks ago: Vicent Caselles. The great friendship and academic collaboration we had is an example of how much I enjoy this work.

I find academia the perfect place to have a great balance between professional and family life. I feel privileged that the flexible time allowed me to pick my kids every early afternoon from school, for over 12 years, while my wife worked. It also allowed me to be at virtually every school event they have, including 9 consecutive years of camping with their school class. My family is my life, and my job allowed me to exercise this. A good friend once told me “the only free people in the world are very rich people and professors, since I can’t be in the first group, I chose to be in the second.” While we work very hard in academia, I believe that much harder than any other job we could do, we work in our own time, and that is priceless.

Yes, there are other great jobs out there, some as exciting as academia, and the world is very different today than it was when I became a professor. Yes, I was very privileged to have outstanding collaborators and students, outstanding program managers for my grants, and incredible support from the administration, and this certainly contributed to my happy life in academia. I always look for other challenges to combine with my current academic activities, other places to contribute and to become excited with. But academia is in my heart, and I hope everybody can enjoy their job as much as I enjoy mine.

I have learned a lot from the Corner Office columns written by my colleagues during the past few months. In order to write about something that has not already been covered, I have decided to focus here on the themes of productivity in academia, industry relations, student mentoring, and international collaborations.

My research is at the intersection of Computer Engineering and micro/nano-systems. To be successful as a researcher in this field, significant collaboration with industry partners is necessary. Therefore, let me first elaborate on the recipes for successful university/industry collaboration.

How to Promote University/Industry Research Collaboration

Research collaboration between university researchers and industry experts requires a number of key ingredients. The first requirement for promoting collaboration is that university researchers work on the right problems. What is a right problem to work on? That depends on the expertise of the researchers, the challenges being faced in companies for which no solutions exist today, and on whether the problem is one of research or development of known technology in a company setting. For example, I have never attempted to collaborate with companies in the area of RF circuit design because my knowledge in this domain is very limited. Also, I have never approached industry partners with proposals for providing ready-to-use tools and software.

Early in my career, I was successful in developing significant industry collaboration in the area of system-on-chip (SOC) design and testing, thanks to the right timing (SOC was the new thing in the mid- and late-90s), research grants from NSF that offered matching funds and strong encouragement for industry collaboration, and interactions with many industry colleagues who could have been university professors with their research mindset. I would travel extensively and give talks at companies, and quite often I would get follow-up emails and phone calls for further discussions. NSF funding was a great facilitator and early career awards from Government funding agencies are always a credibility booster. There is clearly a “winner take all” scenario that prevails in these situations, an outcome no doubt of the emphasis on excellence and merit.

In contrast to the above examples of success, a few collaborations failed to take off when the companies insisted on IP agreements and the university lawyers could not agree with company lawyers. Other attempts at collaboration failed when key people in the companies left for other jobs. Finally, collaborative efforts failed when companies showed no appreciation for theoretical insights and dictated the intellectual content of the research, resulting in frustration and a lack of true collaboration.
In later years, most of my successful collaborations with companies have been a result of outstanding PhD students. Most companies view PhD students as a long-term investment and they appreciate the training provided to them by the university research environment. (Some industry sponsors have even told me with a fair bit of seriousness that they do not care so much for the research that professors do, but they greatly value the training of graduate students.) In my case, for example, my collaborations with companies in board/system-level diagnosis and 3D chip testing have been driven by the hard work of many top-rate students. It is of course incumbent on the Professors in such scenarios to ensure that the companies work with only the best students, those who are motivated, qualified, and willing to make the extra effort to understand industry challenges and practical issues.

Therefore, successful university/collaboration requires a research model that is somewhat different from the traditional university model of “blue sky” research, yet retains the characteristics of independence of thinking and a focus on longer-term problems of tomorrow.

University/Industry Collaboration Models

University researchers must sustain their research programs through external financial support, the right mix of research topics, and high-quality graduate students. In my field (and I believe this is true for many of my colleagues), there is the added burden of ensuring that research is directed towards problems that are relevant to medium- and long-term industry needs, and students are adequately trained for a career in the semiconductor industry. I am faced with a vibrant industry community that includes top-rate researchers and visionaries. Therefore, I have to always ask myself: “How should university researchers engage with this community, and conversely how should industry interact in a meaningful way with the university community?

Several engagement models come to mind as I think about this question:

The oblivious model: Each side deliberately ignores to a large extent the existence of the other, and limits engagement to occasional meetings at conferences and student hiring. This model allows freedom of thought in the most complete sense for a university researcher but deprives the researcher from early access to exciting developments in industry, and it deprives companies from the benefits of harnessing the intellectual prowess and creativity available at universities. Unfettered thinking can lead to major breakthroughs but success stories are rare amidst a large body of mediocre, low-impact research.

The consulting model: A university researcher provides expert opinions and insights on a specific but narrowly defined problem of interest to a company. While such short-term engagements are certainly more meaningful than the oblivious model, they often do not provide benefits to the larger community. Neither do they foster longer-term synergistic relationships. The company dictates the problem to be solved and a university professor completes a “homework assignment”.

The directed research model: Companies often provide research funding to universities in the form of contracts. Many of us in academia have been recipients of such contracts and we have undoubtedly benefited from the support and mentoring provided to graduate students, access to industry data, and guidance on research directions. However, it is often frustrating for university researchers to navigate the hurdles of IP agreements between universities and companies, and the barriers placed on publications and broader dissemination of research findings. In addition, identifying concrete deliverables a priori and meeting them in a timely manner can detract from the intellectual freedom necessary for high-quality research.

The true research collaboration and partnership model: A research collaboration can be most successful and lead to longer-term benefits when the partnership is between equals, research is truly open, problem areas and specific topics are discussed without excessive consideration of commercial benefits for the company, and where mutual respect reigns. Funding in the form of gifts with no IP barriers or burden of deliverables is surely the best way to sustain longer-term collaboration that is beneficial to all. University researchers have the luxury of being able to think in unconventional ways and “out of the box”. Indeed it is incumbent upon them to do so. They can identify and exploit theoretical methods from other fields, as well as form connections between problems that are not obvious to an industry practitioner. In addition to gift funding, such an interaction model is sustained and enriched by the participation of industry mentors in PhD theses committees, industry internships for students, and systematic transition of basic research into industry practice and automation tools. Quite often, such collaborative relationships transcend university/affiliations; collaborations continue in the form of new institutional partnerships when key personnel change jobs.

Irrespective of the collaboration model between university researchers and companies, constant vigilance on both sides is needed to ensure that added value is an outcome of the collaboration, industry sees more value than simply a pipeline for trained graduates, and university researchers find the collaboration to be an enriching intellectual experience. From a university perspective, successful industry collaboration requires a change in mindset—publications should no longer be the only yardstick for measuring success. Publications will always arise from a truly open and success research collaboration, hence the emphasis must be placed on professional relationships that build trust and mutual respect. Making unreasonable demands for funds, data, or resources is counter-productive for a university researcher while the industry collaborator must understand that university research is not expected to lead to ready-to-use design flows or tools. We must start modestly, take small steps, and then go on to bigger things. That is surely a recipe for success.

The Right Research Topics

Academic freedom is a great thing. I am free to work on research problems of my choice, but of course with the caveat that I have to be able to sustain the research program through external grants and the right graduate students. It is tempting to be opportunistic and target research topics that are in vogue, where there is a lot of federal funding, or where there are significant opportunities for collaboration. Such opportunism helps a faculty member in many situations, but the agility with which we change research focus or jump into new areas must be sustained through a genuine appreciation of the area at an intellectual level. When we choose a research topic, we have to ask ourselves whether we will be as excited by this topic after 5+ years, whether our liking for it borders on the obsessive (obsession is sometimes good for research!), and whether we can in reasonable time reach a sufficient level of excellence to compete successfully and be rated as among the very best in the field. I always remember the words of George Bernard Shaw: “Take care to get what you like or you will be forced to like what you get”.

As engineers, we often have to grapple with the need to link new ideas with applications. Conceptual advances in theory or new breakthroughs in technology are often reined in by the need to justify practical applications. Whenever in doubt in such situations, I remind myself of the words of Prof. Herbert Kroemer from the ECE department in UC-Santa Barbara, who won the Nobel Prize for Physics in 2000: “The principal applications of any sufficiently new and innovative technology have always been—and will continue to be—applications created by that technology.” This is also referred to as Kroemer’s Lemma of New Technology! The logic is circular, but so is research and so is the universe in a metaphysical sense.

So how do we start working on a new problem? Our PhD degrees have prepared us to learn new things, and to be prepared to change (as eloquently expressed by Richard Fair in his Corner Office column). First, we have to use our imagination and look beyond the short term. In the words of the French Nobel laureate Andre Gide, “One does not discover new lands without consenting to lose sight of the shore for a very long time“. Also, we have to start “doing” and not just “think” all the time. Finally, we have to be prepared to go against the general consensus and take risks judiciously. I like the pictures below, which express just these thoughts (adapted from originals by Rob Rutenbar, now at University of Illinois).














In my experience, I have found that it helps to have one core area (“home”) where I can maintain a leadership role over a long period of time. With the sure footing of a home base, it becomes easier to feel more secure, and to “roam” and explore newer areas. Research must lead to impact; of course, impact can be measured in different ways, but as a researcher, I need to know in advance how the impact will be measured when I start a new project. There is a certain of sense of speculation involved since we have to select research topics that are just emerging, so that we can ride the wave and be a leader. So this process is a lot like buying stocks.

When I started my academic career, a mentor advised me to always have three bags of ideas. The first bag should contain ideas that could be explained quickly to a practitioner, and through which I could get industry projects and make immediate impact. The second bag should contain ideas for longer-term projects, possibly leading to PhD theses, multi-year grants, and good publications. We should not underestimate the importance of the second bag and revisiting with a critical eye what we already know. Again, I quote Andre Gide: “Everything has been said before, but since nobody listens we have to keep going back and beginning all over again.” The third bag is the most interesting, since it should contain the most high-risk ideas, which might not even be appropriate for graduate student research, might require years of thinking and mulling, but which might one day lead to my magnum opus. I still remember this advice and I make a conscious effort to have all three bags with me!

Another way of looking at research topics is to imagine the figure shown here. My portfolio of research problems at any time includes several points in this space. This strategy has served me well and, again, it reminds me of how we invest money in stocks.

Chakrabarty_2How to Mentor Students

Students are arguably the most important “products” of any university. We do not manufacture artifacts for the marketplace. Instead, we train students for the workplace for a wide range of professions. Hence student mentoring is for me a key job requirement. Here let me dwell on my philosophy of working with graduate students.

Broadly speaking, my role as a mentor for graduate students aspect involves two important tasks: (1) Explain a big problem; (2) Give sound advice and be a trusted guide. I read somewhere long back that a good mentor must be a good listener (key to good communication), be a good problem-solver, and be a good observer (able to spot “problems”).

The first job of an advisor is to dispel fear from the minds of students. When a grad student starts research, we would expect the student to start with a clean slate, i.e., more like a blank sheet of paper on which nothing has been written yet. However, we often overlook the fear mentality that prevents the student from taking the first steps towards leaving his or her imprint on this sheet of paper. In my role as advisor, I take inspiration from India’s Nobel laureate poet Rabindranath Tagore, who penned these famous lines in Gitanjali (the poet’s own translation from Bengali):

Where the mind is without fear and the head is held high;
Where knowledge is free;
Where the world has not been broken up into fragments
By narrow domestic walls;
Where words come out from the depth of truth;
Where tireless striving stretches its arms towards perfection;
Where the clear stream of reason has not lost its way
Into the dreary desert sand of dead habit;
Where the mind is led forward by thee
Into ever-widening thought and action;
Into that heaven of freedom, my Father, let my country awake.

A fearless and proud student is almost always a productive student. All we have to do is to appeal to a student’s pride, and there is hardly a better motivator than the pride that comes from having achieved something notable. A fearless student is never afraid to ask why, question the status quo without being dismissive, and seek answers to difficult questions. In the words of Aristotle, “It is the mark of an educated mind to be able to entertain a thought without accepting it.” Furthermore, students have to understand that success in research requires patience and an inner calm that helps us focus our creative energies (yoga and meditation are of course highly recommended!). When I was new to Duke, Provost Peter Lange told me: “tenure is not an event, it is a process”. Likewise research is not a sprint, it is more akin to a marathon. Again, in the words of Aristotle, “We are what we repeatedly do. Excellence, then, is not an act, but a habit.”

Every student is different, and instead of insisting that every student adjust to my style or personality, I make efforts to understand the unique needs of students and meet them halfway. I have, at various times, provided gratis service (of course with the rider that I am not a professional in these matters!) as a marriage counselor, tax advisor, immigration consultant, mental health counselor, personal safety guide, etc.

International Collaborations

One of the most satisfying aspects of my job is the opportunity to collaborate with outstanding researchers in my field all over the world. The Duke brand name is a privilege, and it is an honor to be a Duke ambassador in these activities. International collaborations offer many benefitstangible and intangible. Tangible benefits include the opportunity to work first-hand with students who ultimately join Duke and work with me for their PhD degrees. This is an example of risk mitigation, whereby I already know a student through a joint project before he or she joins our PhD program. I regularly host visiting researchers from overseas and their presence at Duke for short- to medium-term visits enriches the graduate school experience for my PhD students. Other tangible benefits include participation in joint research proposals (such as EU projects, special programs at NSF, grants from EPRSC-UK and the National Science Council in Taiwan), and an understanding of cultural differences that help me to do a better job in teaching and student advising at Duke. Many companies today are globalized, and US students must understand the challenges and opportunities involved in such a work environment. Researchers from other countries often approach the same problem in different ways, and many of my most satisfying research accomplishments have been inspired by different ways of problem solving. In addition to my Duke family of graduate students (present and graduated), I have a large extended family worldwide that includes students from nearly all the countries that I have visited. Good working relationships with the best researchers all over the world (and an understanding of how they carry out their research) helps me to be a better Editor-in-Chief and Associate Editor of journals.

I have always valued efforts aimed at bridging different cultures, and additional benefits of international collaborations include the opportunity to see amazing places, enjoy exotic food, and learn about history, politics, languages, architecture, lifestyles, and religions. The list is endless.

My colleagues Dan Sorin and Chris Dwyer like to do some good-natured leg pulling. They calculated my average velocity based on all the international travel that I do. Of course, they were technically incorrect; my average velocity is zero since I always return home to Duke. What they actually calculated was my average speed, which admittedly is a really large number.

Final Words of Advice

In my experience in mentoring graduate students and working with colleagues from other countries and cultures, I have learned that I can make a point much more effectively using allegories. So here are three stories with their associated morals (told to me by a friend in Germany).

Lesson 1
A crow was sitting on a tree, doing nothing all day.
A small rabbit saw the crow, and asked him, “Can I also sit like you and do nothing all day long?”
The crow answered: “Sure, why not.”
So, the rabbit sat on the ground below the crow, and rested.
All of a sudden, a fox appeared,
Jumped on the rabbit… and ate it.

Moral of the story: To be sitting and doing nothing, you must be sitting very, very high up!

Lesson 2
A turkey was chatting with a bull.
“I would love to be able to get to the top of that tree,” sighed the turkey, “but I haven’t got the energy.”
Well, why don’t you nibble on some of my droppings?” replied the bull. “They’re packed with nutrients.”
The turkey pecked at a lump of dung and found that it gave him enough strength to reach the first branch of the tree.
The next day, after eating more dung, he reached the second branch.
Finally after a fortnight, there he was proudly perched at the top of the tree
Soon he was spotted by a farmer, who promptly shot the turkey out of the tree.

Moral of the story: Bull*#*@ might get you to the top, but it won’t keep you there.

Lesson 3
A little bird was flying south for the winter.
It was so cold, the bird froze and fell to the ground in a large field.
While it was lying there, a cow came by and dropped some dung on it.
As the frozen bird lay there in the pile of cow dung, it began to realize how warm it was. The dung was actually thawing him out!
He lay there all warm and happy, and soon began to sing for joy.
A passing cat heard the bird singing and came to investigate.
Following the sound, the cat discovered the bird under the pile of cow dung, and promptly dug him out and ate him!

Morals of the story:
1) Not everyone who drops s$*t on you is your enemy.
2) Not everyone who gets you out of s$*t is your friend.
3) And when you are in deep s$*t, keep your mouth shut.

The above lessons can be useful for all of us, as we navigate different roles at different times in our academic career.
In closing, I stress that productivity in an academic career is not possible without a supportive environment. I have been fortunate to have many excellent role models in Duke ECE and an administration that has always been appreciative of my efforts in research, mentoring, industry partnerships, and international collaborations. When I first arrived at Duke, I was under the impression that I am expected to work on a specific problem and in a specific area. I asked Loren Nolte (the department chair at that time) what problem I should work on. I can imagine how much amused Loren must have been, and he still chuckles and reminds me of that episode. I guess I have come a long way since then!

If it had taken me as long to figure out the first words of my first lecture as it has to figure out the first lines of this Corner Office, I am not entirely sure I would have ever made it to my first class sixteen years ago. I am faced now, as I was then, with a new experience, having been given a unique opportunity to talk about my experiences with people in my community. Then, it was all about the software tools I thought would be useful for our undergraduate students – but which were not covered in any one class. Now, it is all about what I think might be useful for teaching our undergraduate students.

I am going to focus on what happens after setting a schedule and getting the right space and teaching assistant support for a course and coming up with a schedule – mainly because I just discovered that writing about those processes took over 1400 words and didn’t really get to the heart of the questions I have been asked to answer! So, now assume you have a course to teach, a place to teach it in, an idea of what needs to happen with the syllabus, and a spherical cow radiating milk isotropically. The latter may not be relevant, but I hope it brought to mind a whimsical image.

The items I want to talk about are lectures, homework, laboratory experiences and assignments, and assessments.


For an established class, deciding what to do with a lecture can either be very easy or very hard. Not quite as hard as deciding when it is time to change research directions – Nan Jokerst, Jeff Glass, and Richard Fair all discussed that particular conundrum in their posts – but still there is very much a sense of attachment to the material in a course and the way it might be conveyed. For example, I still mourn the loss of a lecture on Newton Polynomials that was removed from Computational Methods to make room for a more robust discussion of integration. But with only 35 hours to work with, and having to account for tests, and logistics, and everything else, there is only so much that can go into a course. I know for myself that I have to remain vigilant that I do not let something that is personally interesting – but which does not provide the students with an appreciably or proportionally greater understanding of the material – take time away from other things.

My notebooks for classes have day-by-day dividers in them so I can put the notes and references for a particular day in their own place. I also generally try to use a Sympodium or a slate PC for writing “on the board.” This does a number of things for me: I can use many, many colors; I can go back to previous slides if someone needs to see them; and I can print out what I wrote on a given day and put it in the notebook. I thus have a history of the semester’s material that I can use to improve things in the next cycle.

There are some disadvantages to using a Sympodium. First, you are stuck standing in one place – and if you are in Teer 203 or Schiciano, that one place may be relatively far from where the projection screen is. Second, you are beholden to the resolution of the screen for how much material you put up at once. Third, when you point to something and say, “look here,” no one has any idea what you are talking about unless you make some obvious motion with the cursor or use a laser pointer. Or, in Teer 203, come out from behind what I call the Wizard of Oz Box, walk on stage, and actually point to something.

For a new class, deciding what material to be included and how to parse it out into lectures can be very difficult. For Mechatronics, I made a list of the various topics I wanted to cover, wrote lecture notes on my tablet for them as one continuous file, and then began cutting and pasting the material I got through on a particular day. I then looked at how the course coverage was matching the specific items listed for each day on the syllabus and adjusted. As it happens, a few weeks into that class I realized that the pacing and content thus far were not working; I ended up taking a few lectures to go back over some territory and re-mapped the syllabus for the rest of the semester.

Regardless, I always try to work in a couple places where the students can work together on some aspect of the course we have discussed. For introductory classes, that might mean finding equations that will model a circuit or writing a bit of MATLAB code; for upper level classes, it might mean coming up with a plan to implement a particular controller. Whatever it is, the process of engaging the students with the material and then hearing back their answers will serve two purposes. First, it will get them talking – which always helps if a few of them are a bit sleep-deprived. Second, it will give you a little bit of instant information as to how the material is coming across. There may be groups who cannot come up with correct answers or who ask clarifying questions while they are working – finding that out during the lecture, when there is still time to provide some form of clarification to the whole class, can be very valuable.

I also try to stay on top of campus news to make sure I know what is going on with our students in general. A quick read of The Chronicle (Duke’s, not “of Higher Education”) can be very valuable in learning what events our students see as important. Including those items in a lecture can also help reinforce the notion that We Are Duke. This can take place in the minutes just before class starts as people are filing in or even during a lecture if it is relevant.
Mainly though, be sure to take Nan’s advice on teaching: “Be organized and enthusiastic, teach clearly, and treat the students with respect, and you will do well as a teacher.” (ECE Corner Office by Nan Jokerst, 02 Nov 2012)

Effectiveness of Homework Assignments

I have taught a fairly broad range of courses – everything from Introduction to Engineering Computer Programs to Seapower & Maritime Affairs – and for each I have had to determine the function of homework. There have been some common components:

–  Homework is most effective when the feedback loop is accurate and efficient. Students should get back thoroughly examined material in a reasonably short period of time. I have always worked to get the number of teaching assistants that will meet this need. For the classes I teach, I also provide TAs with a set of solutions and a grading rubric. If possible, I will have assignments split into parts and I will have one TA grade all submissions for a part. In ECE 110 I have two grader TAs, so students turn in a Part I and a Part II for each homework. All 40 people will have their Part I graded by the same person, which hopefully leads to more consistent grading. There is a bit of overhead involved – two entries in the gradebook and two piles to collect and redistribute – but the overall process has worked well.

–  Homework is most effective when there is a means of comparing the finished work against legitimate processes. This does not necessarily mean “a correct solution,” because oftentimes work done at home will and should provide open-ended opportunities for students to develop their answer in a way that is unique and is in keeping with the principles established by the course. Doubtless, there are some assignments where there will be A Right Answer – but generally there are multiple different paths to get there.

–  Homework is most effective when students have the tools to complete it. An assignment does not need to look exactly like an example from the book or from a lecture, but there should be some reasonable expectation that the underlying information and skills required to solve the problem are available to the student given a reasonable amount of work. Which, of course, means defining “reasonable.” Any takers?

One part of how I use homework which varies from class to class is the notion of collaboration. This is a complex issue. There is value in requiring that an individual, alone, gathers all the information necessary to develop a solution to a given prompt and develops that solution alone. There is value in having that same process be shared by a group with all the viewpoints that may be discussed along the way. What is critical, however, is giving clear guidance about collaboration policy. Gary Ybarra and Lisa Huettel, for example, have crafted the controlling document for ECE 110. The end result establishes what is and is not allowed, gives examples, and even informs the students of historical repercussions for failure to abide by the rules. It also notes that there is a “gray area” and encourages students to contact instructors in those cases.

Laboratory Experiences and Assignments

For those classes with hands-on labs, the experience and insight gained from performing the work, analyzing the results, and crafting the conclusions can be a very powerful component of the course. For the Computational Methods course in particular, labs are the primary means by which students will actually learn the material. For other courses, such as Fundamentals of ECE, labs are the avenue students take to become familiar with building blocks of ECE. The most effective laboratory experiences I have seen have some common components as well:

–  Laboratory experiences are most effective when they complement or clearly supplement lectures and readings. If a lab involves material that has not been covered yet, students can either miss the point or get frustrated at not being able to understand what is going on. If a lab is going to introduce some new concept, it is very important that the period during which the lab is performed by prefaced with a period of instruction on that material. That may involve a class lecture, an in-lab talk from a laboratory TA, or clear reading material – preferably with examples. If the lectures for a course start slipping behind for some reason, consider re-working the lab schedule or determine a means by which the students will be otherwise prepared to truly learn from the laboratory experience.

–  Laboratory experiences are most effective when there is an assignment requiring that the student takes information from the lectures, from the readings, and from the laboratory itself and combines them to demonstrate a broader understanding of the material as a result of having done the lab.

–  Finally, laboratory experiences are most effective when they work. Admittedly, learning the mechanisms by which an experiment will not succeed can be very important, but for early classes where students do not have sufficient troubleshooting experience or theoretical knowledge to “fix” a malfunctioning lab setup, care must be taken to create a laboratory experience that provides legitimate data and supports the theories discussed in class. Also, whomever runs the lab must be prepared as well.


This section of my post will not be on the test. But it will be on tests. In many courses, especially early ones, a healthy portion of the assessment of a student’s learning in a course will be done via tests. And, depending on the course, of the 35 hours devoted to a standard 1-credit course, tests generally take between 0.83 hours and 1.875 hours total depending on if you have two or three tests and whether you teach in 50-minute or 75-minute sections. That either represents a small or large portion of the whole depending on your perspective. In my experience:

–  Tests are most effective when they can be completed in the time allotted. I know there will be some disagreement here, and my experience is biased by the nature of the courses I teach. I also want to emphasize the “can” here – there is certainly no requirement that every test be written such that every student in a course can complete it. But there does need to be a sense that a well-prepared student can complete the tasks at hand in the time allotted. For some of my classes, this has meant moving the test to the lab, where the students could use 170 minutes to work on the assessment. I will say that this is one area I continue to struggle with – as much as I try to account for my experiences as well as the fact that it will not take me any additional time to interpret a problem (since I wrote it), I still get this wrong.

–  Tests are most effective when they efficiently cover the breadth and depth of material. I will generally give a class a “road map” for an upcoming test. For Computational Methods, the first road map is a 5-page long outline of concepts and commands totally 23 topics with several sub-topics. The second is a 1-page long outline with 6 topics (including “Know everything from the first test”). Once I make that map, I work to have some aspect of each part show up somewhere on the test.

–  Tests are most effective when they ask valid questions. I have managed to give some test questions that were either impossible or trivial. Allow me to illustrate from ECE 61L Test III Fall of 2001:

Gus' Corner Office Image

Just before I went to copy the test, I decided to be clever and ask for the current through resistor R3 instead of the inductor current. Hilarity ensued. Almost everyone got a 0 for that problem, which means almost everyone got a perfect score on that problem.

Given the pressure that students feel during a test, and given the contribution of the testing process (studying for, taking, and getting back) to the learning process, it is crucial that tests be legitimate.

Lessons Learned

I’ll cap this off with what I think I have learned from the experiences I have had teaching so far.

–  Keeping things organized is essential, especially for larger classes. For me, that means having a spreadsheet with all my to-do items for each of my classes as well as structures in place for exchanging materials with the TAs and returning materials to the students. For my Mechatronics class, as an example, I have a rolling cart with 80 hanging folders – one per student. TAs return all graded material to each student’s folder, and then I roll the folder to class or lab for recovery. It greatly reduces the time it takes to have students find their stuff and gives me an easy visual way to see if people have been showing up to class.

–  Keeping track of “where we are” is crucial. I generally teach Monday and Friday, and a great deal goes on between lectures. Be sure to note for yourself where you actually stopped in a given lecture, not what the schedule said you were supposed to cover.

–  Knowing where things fit into the grand curricular scheme of things can really help. I have a head start here comprised of having gone through Duke as an undergraduate, of having had appointments in three departments, and of having taught classes for all four. Since I teach or run lab for all incoming engineers, knowing the different degree requirements for all of Pratt is a job requirement. Knowing what goes on for all of ECE on the undergraduate side – not just within on curricular group – can help you include connections between the material you are teaching and the material in other classes. That, in turn, can give the students a better sense of the big picture.

–  “Cooperation Without Compromise” – I first heard this as the motto of the Navy Chaplain Corps. As an aside – it is decidedly more appropriate here than the motto of the part of the Navy I was in, the Seabees, which is Construimus Batuimus (“We Build, We Fight”).

The idea is that there may be times you need to work to cooperate with a student or group of students without compromising the integrity of the course or your own integrity. For example, some of our varsity athletes have rigid travel schedules – working with them in advance to overcome those obstacles may involve moving deadlines or providing other accommodations. Other students will have life events that merit consideration. And that consideration takes time and effort on your part, but it will be worth it.

–  Be approachable. Come to class a little early and stay a little late if students want to bend your ear. For your office hours, emphasize that the main purpose is to discuss concepts from the course, but that conversations on other topics are certainly welcome too.

If any part of this has been useful, there are many members of the staff, faculty, alumni, and students who should get the credit. The rambling, however, is on me. Duke in general and Pratt specifically have been my home now for nearly 24 years, and I wouldn’t have it any other way. I consider it a very real privilege indeed to be a member of the ECE faculty, and I’d like to thank Larry for the opportunity to contribute this post. I have learned much from the posts that have been in the series so far and look forward to reading more.