The next generation of robots will complement humans by being more efficient, more resourceful, and better integrated into our society in performing daily tasks.
Devesh Jha, our next pathbreaker, Principal Research Scientist at Mitsubishi Electric Research Labs (MERL, USA), works on problems related to robotic manipulation in an unstructured environment.
Devesh talks to Shyam Krishnamurthy from The Interview Portal about the huge potential of robotics in a world where robots will coexist in the same space as humans.
For students, cultivate a mindset of innovation, because when you focus on innovating, obstacles become opportunities!
Devesh, can you tell us about your growing up years?
I grew up in Patna, in a modest, middle class family. My father used to work for Bihar State Electricity Board (BSEB), and my mother is a housewife. My father is retired now. I have three elder siblings. So, with four siblings, we had a big family. I studied in the local DAV school which came to be known as DAV BSEB later in Patna. We used to live in the state government housing provided by BSEB to my father. As such, I had a lot of neighbors who were also my friends. Some of my best memories growing up were sharing our school lunches, the sunday breakfast where I would stuff myself with Aloo paratha (made by my mother), and the hide and seek games played with my friends during summer while there was a electric load-shedding in our neighborhood. One of my fondest memories growing up was going to my grandparent’s place during summer vacation. My grandparents used to live in a remote village in Madhubani—that is where my father grew up. My mother also grew up in a neighboring village. During summer vacations, we would go to our grandparent’s place where all our cousins would gather as well. That was a lot of fun—life used to be very simple those days. I would be happy just going to the local religious pujas in our village, eating the prasadam from the puja, and meeting my cousins. We did not even have electricity in our village when we were growing up, but yet it was so much fun. I think I was fortunate to have been exposed to such a simple lifestyle as that keeps you grounded.
I think I was always someone who was dreamy, and could get motivated very easily. That being said, I could never understand why I was studying what I was studying—that makes studies less interesting (I think this is also something that we need to work on in our schooling system). In middle school, I got interested in Math and History. My interest in math continued in High school but unfortunately, I could not keep up with my interest in History. I believe this was mostly due to the pressure of performing well in High school so that I could get into a good engineering school.
I grew up playing gully cricket. Like most of the cricket fans, I was an avid fan of Tendulkar. For me, cricket was about him. Tendulkar was my idol while I was growing up, and I would draw inspiration from his dedication and efforts. I also used to play a bit of soccer and badminton depending on the weather and group of friends. I also drew a lot of inspiration from my elder brothers who went to engineering colleges and were always considered smart in my parent’s friend circle. Honestly speaking, in school, my inspiration was limited to just becoming an engineer, and probably going to a business school afterwards (the stereotypical middle class mentality). I understand where that comes from now—at that point, I neither had the understanding nor the exposure to think about anything beyond that.
I was able to score decent grades in my 10th board exams, and then decided to go to the Kota factory to get in the race of IIT-JEE. Looking back, I think that was a bad decision as I never got to enjoy my junior and senior years of High School, and somehow I ended up putting too much pressure on myself to crack JEE. Based on my experience so far, I would like to encourage students to understand that one exam just cannot make or break their fate in life. Unfortunately, we still have a traditional way of judging success in our society (or at least in the middle class society). I think getting into one of the IITs does provide that initial, quick path to success (or at least the feeling of success). I sincerely hope that this changes over the years, and our coming generations are taught to innovate, and not to go through a two-year training program to get into engineering colleges.
What did you do for graduation/post graduation?
I went to Jadavpur University, Kolkata for my undergraduate studies in Mechanical Engineering. While I was devastated that I could not make it to any of the IITs, I was fortunate to make it to JU. My time at JU was definitely one of the best times I had in my life so far. It was also an exciting phase of my life as I was trying to figure out the next steps in my life, and where I would go in the future. I made amazing friends during my undergrad at JU.
In my initial years during my undergrad, I was in college without any purpose. After meeting some of my mentors in junior and senior year at JU, it somewhat became clear about what I would do in my career. I developed an interest in some of the engineering subjects I came across, and I developed a curiosity. Simultaneously, I also developed an urge to pursue innovation in what I was doing. This curiosity and inclination to innovate led to my decision of moving to the US for a PhD.
What made you choose such an offbeat, unconventional and unique career?
It started in my family. My elder brothers are all engineers, so that was kind of what I ended up doing as well. But then some of the professors in my undergrad curriculum (as I mentioned above) influenced me to pursue a PhD.
Apart from meeting friends, I also met some of the best teachers in my life at JU. These are some of the brightest people I had ever come across in my life so far. This was also because of my mentors, Dr. Sumanta Neogi and Dr. Sandip Das at JU. Being a faculty at these universities, while fighting the system provided to them to inculcate innovation, is an extremely challenging job (to say the least). However, my mentors still had the drive to help students and guide them to achieve their best in their career. That is an extremely challenging job and this experience of working with them always motivated me to keep going.
I was doing well in my studies at JU, and my good grades at JU made it easier to get admission at Penn State into their PhD program and receive full assistantship from my PhD adviser. However, I had no idea of the US schooling system, and the lifestyle there. I also had a good job offer at that moment, and my parents wanted me to take that job, and maybe, make it to a good B-school later (things that are common in all middle class families). I think at that point I did not realize how fortunate I was to get this opportunity to pursue my dream of innovating for a living. I can see that now, it’s very hard for students from modest, middle class backgrounds to choose career paths which are non-traditional. Now, this is also a cause that is very close to me, and I would like to contribute to change this.
At this point, I met my PhD adviser, Dr. Asok Ray – I think this was a big turning point in my life and career. Dr. Ray is an immensely smart and enlightened researcher, a wonderful mentor, and a selfless human being. I have not met many people who share the same level of drive and enthusiasm as Dr. Ray (especially at his age, he received his PhD in 1976). At Penn State, I did my PhD in Mechanical Engineering while also pursuing Masters degrees in Mathematics and Mechanical Engineering. I cannot thank my adviser, Dr. Ray, enough, in pushing me to pursue my Masters degree in Mathematics. While it was difficult, I met some of the smartest, driven people one can meet. The drive, innovative thinking and energy of these people had an extremely positive influence on my life.
While my PhD is in Mechanical Engineering, I wanted to pursue a career in robotics. This was a bit misaligned with what my adviser’s research thrust was. However, my adviser always encouraged me to pursue my own research interests and I ended up working on problems in autonomous driving, machine learning, etc. for my PhD thesis.
My PhD adviser, and my lab mates at Penn State had a big effect on my critical thinking and independent thinking. They always made sure that I did not get comfortable, and somehow these habits and instincts stuck with me. Moreover, the challenges along the way to getting my PhD always kept pushing me to try harder.
When I started working at MERL (Mitsubishi Electric Research Labs), I was fortunate to get really good colleagues and collaborators. They are some of the best brains that one can interact with. I found new mentors, and kept moving toward more challenging problems in my field. This has been another key influence on my career trajectory so far.
Tell us about your career path
I went to pursue my PhD right after college. I got full assistance from Penn State for my PhD. During my PhD, I did an internship at MERL which was my first exposure to working on real world problems. It was definitely an eye opener for me. I was working on algorithms for autonomous parking of cars, which was as exciting as it sounds.
My PhD thesis tried to look into feedback control of autonomous systems using machine learning and planning/control techniques. This mainly involved making inferences out of data, and using these inferences to control the autonomous systems.
Most of the challenging problems that one comes across in AI/robotics have a deep mathematical part which very few people understand. My PhD adviser understood this, and he asked me to consider doing a specialization in Mathematics. Mathematics is a language, doing a specialization in Math will help you express and understand the engineering problems better, especially the things you cannot express or understand in English (or any other language). Also, it develops your critical thinking, and one benefits from it no matter which field they end up in.
During that time, Artificial intelligence and robotic manipulation were also getting a lot of attention. I was looking for an opportunity to pursue that topic. Since the nature of my job is more technical, my strategy at that time was to publish papers in the top AI conferences so that I could be considered for such positions. I was fortunate enough that my current employer had an opening on this topic and I was selected to work on it. Since then it has been a very challenging but very rewarding journey.
How did you get your first break?
I got an internship at MERL where I was working on autonomous driving problems. That was critical for me to get my current job at MERL. This gave me the unique opportunity to work on exactly the topic I wanted to work on after my PhD.
What were some of the challenges you faced? How did you address them?
If you are trying to do something meaningful, there are always challenges. I did not face any particular challenge as such. But my current role is very challenging technically, and always keeps me on my toes. But, then I found help along the way. I found really good mentors, collaborators, and colleagues which make these challenges easier.
Where do you work now? Can you tell us what you do at Mitsubishi Electric Research Labs?
Currently I am a Principal Research Scientist at Mitsubishi Electric Research Labs (MERL) in Cambridge, MA USA. I work on all problems related to robotic manipulation. My research is based on the theme of “robotic manipulation in an unstructured environment”, which means I am trying to make robots coexist in the same space as humans, where they can be more intelligent and dexterous. My job is very technical, and hands-on. I am fortunate that I get to decide and then work on the research problems that I think are important. That is something which is unique about my current job. However, this does require lots of skills. Since I work in robotics, a key requirement is the experience to work with robots. It also requires deep knowledge of optimization, machine learning, mechanics, and artificial intelligence.
A typical day could involve working on mathematics of a new algorithm, or writing a research paper/patent draft, or testing a new algorithm on a robot. There’s never a dull moment.
If interested, you can check out my profile and a few links to my work at MERL:
I love robotics because it is not only intellectually challenging but also because everyone loves robots. Nothing is cooler than robots making a smart move. So, it makes it gratifying at the same time.
How does your work benefit society?
I hope that with artificial intelligence, the next generation of robots will be more efficient, and this would lead to a society which is sustainable, by optimizing the use of resources. Humans will do the most noble jobs like teaching, medicine, etc.
Tell us an example of a specific memorable work you did that is very close to you!
I was part of a work that was displayed at a press release at Mitsubishi Electric Corporate Headquarters in Tokyo, Japan. Going through that process made me more humble, and gave me an idea of the effort we have to make to go from an idea, to the media, and then finally to a product.
Your advice to students based on your experience?
I think I would ask them to focus their effort on innovation. We can only get ahead by innovating. Countries like the US are ahead of everyone else because innovation is part of their lifestyle. We need our students to focus on innovation. We, as a country, need to understand that others are only going to take us seriously as long as we keep innovating. Right now, India is just viewed as a huge market and source of cheap labor. We need to shift this view and need to be known for innovation. Of course, this needs a lot of support through better policies from the Government, but students should also develop an appetite for innovation.
My dream is to democratize robotics—this is something that I have been thinking about a lot. We have all witnessed the AI revolution which has been brought up by democratization of machine learning (meaning these techniques have been made easily available, they are open-source, etc). Robotics, on the other hand, remains unreachable to a lot of students, especially in India. I have this dream of democratizing robotics, where students can get exposure to robots even in high schools through different special camps. This is the time when robotics is approaching another inflection point where robots are getting better and better. Soon they will be available as home assistants. They are going to clean our dishes and assemble our furniture. There is a huge potential, and we need to educate students so that we are ready for the oncoming robotics revolution.
My own experience of the struggle that one has to go through just to pursue an unorthodox career path while coming from a modest, middle-class family background also gives me a big reason to help and mentor students at an early age. I do believe that this path could be very much simplified if students are provided the right guidance at the right time. This is something that is very easily available to students in some of the best universities (and that is why we see them succeed). This needs to change for students in India. We need to mentor our students better, so that they can bring innovative ways of changing our society for the better. I am looking for opportunities to mentor more and more students, and help them shape their critical and technical thinking as well as making the right career decisions. I can tell you that there is nothing more gratifying than shaping someone’s career and helping them get where they can get.