The future of electric power systems will be centered around delivery of more efficient, optimized and good quality power through intelligent distribution networks.
Utkarsh Singh, our next pathbreaker, Senior R&D Scientist at depsys SA (Switzerland), works on smart grid monitoring technology, to ensure a reliable distribution grid network using algorithms.
Utkarsh talks to Shyam Krishnamurthy from The Interview Portal about the tremendous potential of multidisciplinary fields encompassing electrical engineering, signal processing and machine learning in finding the root-cause of power quality problems in the realm of energy transition and new technologies.
For students, you have the opportunity to spearhead our energy transition efforts, by ensuring reliability in the conventional grids, which are slowly transitioning towards fully-connected networks or standalone operations based on renewables.
Utkarsh, Your background?
Hello readers, my name is Utkarsh and I was born in Varanasi on 17th September 1990. I completed my initial schooling in Varanasi and Prayagraj. In 1996, we moved to Roorkee where I completed the rest of my schooling in science. I had opted for the combination of Mathematics and Computer Science in 11th and 12th standard. I chose these subjects because I wanted to pursue Engineering and I had a keen interest in programming & logic. Physics, Chemistry and English were obligatory. I always had a bit of struggle with Math, but persistent efforts helped me get good marks.
My father is in the teaching profession and my mother is a homemaker. I have a younger brother who is an Electrical Engineer in the Modern Coach Factory of Indian Railways.
My early interests were in sketching, poetry writing, playing cricket and football. Now, I spend my free time listening to music, reading about technology, geopolitics, history, culture, and human behavior. I also like to go for long walks.
What did you do for graduation/post-graduation?
For graduation, I couldn’t make it to the top-tier Universities. I wasn’t a very bright student though I was quite hard working. I completed my graduation from Uttar Pradesh Technical University in Electrical & Electronics Engineering between 2008-12. I did my Master of Engineering in Power Systems (2012-14) from Thapar University. After my master’s, I did my PhD (Power Quality) at IIT Roorkee.
Can you tell us about the influencing factors that led you to such an offbeat, unconventional and uncommon career?
I was initially considering Electrical and Computer Science Engineering as my options for my BTech. My final choice was based on my inclination towards Physics (especially electrical circuits). I enjoyed coding, but I was not a very efficient programmer in those days.
I wanted to do a post-graduation with a specialization in power systems. So I started preparing for GATE from 3rd year. Although I passed GATE by a narrow margin (as the exam was quite tough), for the first time I felt a sense of achievement and motivation to perform better. In the final year I finished my degree with 80% (Hons.) and also managed to significantly improve my GATE score. I also managed to get 2 placements, but since my future prospects were not very good, I went for a Master’s immediately after graduation.
I got enrolled in Thapar University for a Master of Engineering in Power Systems (2012-14). Based on my GATE score, I got a scholarship. I strengthened my coding/simulation, teaching, and conversational skills during that phase. I actively took responsibility for various group tasks/ leadership roles.
After Master’s I had a choice between PhD in IIT Roorkee and Infosys. I went for my PhD and finished my research work in almost 3 years. My PhD research work was based on the detection and classification of power quality disturbances, where you have to detect short term disturbances (like sag, swell, etc.) using signal processing techniques. We need to further classify using feature extraction which can help in identifying those differences (for example a rise in voltage signal amplitude). Training machine learning algorithms on such magnitude and frequency based features can help in automatic identification of such disturbances and take timely mitigative actions. Thus, I got tremendous experience in multidisciplinary fields like power quality, signal processing, machine learning and optimization.
I also found time to volunteer for several technical events. Patience, persistence, and exercise helped me maintain my sanity through the tough times. Never compromise on perfection, output matters! That was another takeaway.
How did you plan the steps to get into the career you wanted? Or how did you make a transition to a new career?
My objective was to be in research in the field of Power Systems. To ensure that, I pursued a specialization in the same. I didn’t leave any opportunity to learn about practical aspects of the operation of power systems. I used to converse with my professors, industrial acquaintances and work on lab set-ups in the University. Not getting a job aligned with my field was yet another reason to move ahead and pursue a PhD in Electrical Systems. I enriched my skills not only in power systems, but also in signal processing and machine learning (ML). I also acquired hardware skills in real-time digital simulations. Researchers who have multidisciplinary knowledge, especially in machine learning, are currently in huge demand. That is how I landed my first job offer at Université libre de Bruxelles (ULB), where I worked on deep learning algorithms applied to communication systems for crowd monitoring.
The Crowd monitoring work was quite different from power systems research. However, a background in signal systems and applied machine learning proved to be a useful set of skills which got me this job. Also, a number of publications in top-tier journals helped me prove my research ability. In this project our task was to use wireless sensor networks that ping on smartphones in an area, to estimate the number of individuals present in that area (and thereby the crowd density). Based on the collected data, we developed and trained deep learning based forecasting algorithms to predict the number of people that would be present in a given area in a given interval of time. This, of course helps the event organizers and security professionals in being prepared for managing the crowd for ensuring safety.
The work experience at ULB got me another job in the Artificial Intelligence lab in Vrije Universiteit, Brussels. I could learn much more about using AI and contributing for the common good.
In the AI lab I was involved in several technical and non-technical projects. One of the most interesting and challenging tasks that I contributed to was, preparing training material to introduce AI to non-native people. It was interesting to brainstorm ways to explain the pros and cons, how AI is an expert system and a stupid bot at the same time. We designed a simple game based on an AI algorithm to explain all such aspects to people who would go through the training. This is indeed the only way one can explain to people why AI is important (various useful applications) and not a threat to humans (it won’t fully replace them, because it often needs human intervention). I also contributed in writing various research proposals for funding based on AI applications in healthcare, food technology, cyber security, etc. I was amazed by the farsighted approach of the lab.
I loved the challenges and opportunities in that job, but had to move on and join my current job where I was being offered a Marie-Curie fellowship for an industrial project.
I had jointly written the project with my present company and applied for a Marie-Curie Individual Fellowship. It is interesting to mention that before this I also had experiences of failure where I had applied for this fellowship with 2 other labs. Once immediately after finishing my PhD, in a Polish research lab. The second time, I had applied for the fellowship at ULB with my then supervisor, as an extension of the crowd monitoring project. The rejections should only be seen as a learning opportunity. Each failure teaches you how to improve and gives you more confidence in writing project proposals.
It was a great achievement in itself and a good opportunity to head back to the domain of power systems research. I had worked very hard on writing the project. Though I had missed the funding by a small margin in first attempt, my project became one of the top funded projects in the next attempt. In the present job, I am involved in Smart Grid technologies, energy transition, with special emphasis on Power Quality. I help distribution system operators understand why power quality is important for them and what they can do to ensure reliable and efficient operation of their networks in the age of energy transition.
When I say networks, I am implying that the developed algorithms are capable of working on both fossil fuel and renewable powered power system networks. That is how we can ensure energy transition, by deploying monitoring algorithms which can ensure reliability in the conventional grids, which are slowly transitioning towards fully-connected or standalone operation based on renewables.
Excellent earning and networking environment were the key factors which influenced me in choosing my present career abroad.
I always stayed in contact with the people in my field. I was always seeking advice from my professors and reaching out to the institute alumni network for mentorship.
I would also like to mention that I always have my goals in my mind, but keep a flexible attitude. I trust my capabilities, hard work, and ability to get along with people easily. Because of that I can do well and thrive in any environment.
How did you get your first break?
My first postdoc opportunity was in an applied machine learning role in the OPERA-Wireless Communications Group in Université libre de Bruxelles in Belgium. I had submitted my PhD thesis in October 2017. I had decided that I wanted to have further research exposure in an international environment. After my thesis submission, I immediately started applying for postdoc opportunities abroad. I had around 40 applications in various labs. The applications included contacting Professors personally, applying to specific projects and writing joint project proposals with Professors for fellowships abroad. By December 2017, I got reports for my thesis evaluation which were all good and by February 2018, I was able to defend my PhD after the availability of my external examiner. By April 2018, I had 4 confirmed postdoc offers in hand. 2 of them specifically in power systems research, 1 of them in brain computer interfacing based on my signal processing experience and 1 in crowd monitoring at ULB due to my machine learning experience. After a lot of thinking and consultation from various professionals I decided to choose the opportunity in ULB. Working in Brussels could give me the exposure that I was looking for, and it was also an opportunity for me to establish myself in further in the field of machine learning and artificial intelligence.
The advice from various professionals in the field turned out to be very crucial in my choice of a Postdoc. There were also countless bottlenecks in terms of official procedures which were eventually sorted out.
What were some of the challenges you faced? How did you address them?
I do not keep a point-wise list of challenges. Just remember your struggle and capabilities and learn from your struggles to come out stronger and more capable in the future.
Where do you work now? Tell us about your work in Power Systems
I have been involved in Research and Development. Currently, I am working as a Senior Scientist in the R&D department of depsys SA – a Swiss company working towards smart grid monitoring and energy transition.
I am working on smart grid monitoring technology, especially on the problems centered around power quality localization and improvement. My work is to analyze power quality data from various customer grids and help the network operators in maintaining good power quality. This helps in ensuring a satisfied customer and reliable distribution grid.
What are the skills required in your role?
This job requires strong research skills and analytical capability. On a finer level, the knowledge of power systems, power quality issues, vectors, statistics, and coding is needed. Good communication skills are also needed because I often need to present complex scientific topics from a layman perspective. The customers must be able to understand why a problem has arisen, why it is necessary to address it, how it can be addressed and how to maintain the performance in the long-term.
A typical day at work starts with a short meeting with the research teammates, where we briefly discuss personal well-being and professional goals of the day. The objective is also to see if we could assist each other in solving any bottlenecks in our tasks. Rest of the day is mostly about working towards research objectives, short-term company goals, and sometimes additional meetings on various topics.
What do you love about your job?
I love this job because it offers a perfect mix of challenges, collaborations, opportunity to work on and contribute to state-of-the-art monitoring technology, and learning opportunities. It really provides an insight into what actually matters, how things work, and what is going on in the energy market.
How does your work benefit society?
My work is centered around delivering more efficient and reliable power distribution networks. By helping the network operators deliver good quality power supply, we can indirectly ensure thousands of satisfied customers by just using a few algorithms. Thus, my work is essentially directed towards the society at large. Of course, a lot of hard work goes into developing such algorithms and codes. My algorithms are focused on finding the root-cause of power quality problems in the realm of energy transition and new technologies.
Tell us an example of a specific memorable work you did that is very close to you!
I remember that I was assigned the task of data and root-cause analysis for power quality issues in a customer grid, within 3 months of joining. When you are entrusted with such important and crucial tasks at such an early stage, it not only gives you an opportunity to learn while doing, but also a sense of responsibility and appreciation.
I managed to complete the task very well and showed the value of our product and analytics to the customer. It led to huge improvements in system operation and also retaining the customer after the trial phase. I also got the opportunity to share the results with the customer success team and the senior management. It was a special experience for me.
Your advice to students based on your experience?
I have a set of points for the young ones based on my own experiences:
- Work hard on your own, do not have the herd mindset. For example, tuition and coaching are not always helpful.
- Be persistent in your efforts, do not give up easily. When tired, take a break. Practice meditation and exercise on a daily basis.
- Always have a good network. Personal and professional mentors can help you immensely. Find the right people from whom you can learn something new and receive the right guidance.
- Ignore any hate or jealousy. Also, do not be jealous of others. Always use your energy in a positive way.
- Read, converse or watch – to learn about things that catch your interest. Become an expert in those topics. The more you do it, the more knowledgeable you will be over time.
- Always keep a backup when looking/ applying for opportunities.
- Sometimes things do not work, it doesn’t matter how hard you try. Be patient and prepared, devote yourself to whatever comes your way, and you shall achieve success at the end.
- Do not be afraid to learn new things or work in new domains.
- Spend time to understand your own personality and work towards improving interpersonal/ communication skills.
- Keep learning and keep networking!
I would like to give the mantra I received from one of my supervisors – ‘In order to progress, you should know your worth in terms of every second and you should work to continuously increase your per second value’.
Overall, the best way to live is to enjoy the present. Do not overthink the past or worry about the future. Always have plans, but never reveal them to everyone 😉