Yaman Kumar Singla works as a research scientist in Adobe's Media & Data Science Research (MDSR) lab.

Yaman is a young researcher at the Heidelberg Laureate Forum and has also won the Prime Minister PhD Fellowship (by the GOI). The Science and Engineering Research Board, Department of Science and Technology, Government of India, runs a well-known program called the Prime Minister's Fellowship for Doctoral Research. This program aims to improve academic research so that it meets the needs of industry. 

INDIAai interviewed Yaman to get his perspective on AI.

In your opinion, what are the main differences between software development and research? How did you deal with your transition from a developer to a research scientist? 

Both software developers and researchers are creative problem solvers. The sheer amount of change you can see in the last 30 years can be attributed to the technological revolution. Just a few decades back, one could not imagine sending messages and payments instantly and sharing one's views through mediums such as this blog. All of this has been possible because of developers and scientists.  

I have been fortunate to be a part of both communities, and each one of them is impressive! I feel that the primary difference is how they define their problems. While research is about finding an answer to slightly open-ended questions that may not have an answer, good engineering is about building solutions and making them work. A researcher is happy to discover something new, while a developer is happy to see his creation in the hands of users.  

In my case, I never left research. Therefore, even working in a software development team, I collaborated with a few like-minded folks to discuss and implement tech-enabled solutions like recognizing speech from silent or noisy videos during my free time. So, transitioning from a development role to a research role was not entirely new. Similarly, my new role as a research scientist also involves building pipelines which can help users access features enabled by my research.    

I recommend everyone working on either "side" to experience both and see each role's unique challenges. 

How do you deal with the most challenging times in your research? How do you keep yourself motivated? Who do you look up to as a role model? 

I am a follower of Advaita Vedanta. In his book Karma Yoga, Swami Vivekananda said that work allows one to achieve the same ideals as Buddha achieved through meditation and Christ through prayer. Work, according to Vedanta, is a way of service and an expression of love. I always feel privileged to be in a field that allows me to do something for the world. While there are challenges in one's day to day, solving those challenges is part of what helps one to achieve those ideals.

Moreover, solving the challenges is what makes my field so satisfying. They help me hypothesize, experiment, and in a way, prove (or disprove) the hypothesis. In addition, the joy of discovering something new helps overcome any temporary stress in the day.  

What are the unique opportunities of doing PhD in India? Can you explain your dual (Joint) PhD from the State University of New York at Buffalo and the Indraprastha Institute of Information Technology in Delhi?  

Yes, first, I'll explain what a joint PhD is. Quite simply, one enrols in two institutes, has two advisors, spends time at both universities, and finally earns a degree containing both names. Why do a joint PhD at all? Well, there are many reasons: 

• Two different advisors with each bringing their expertise and ideas 
• Two different research cultures 

Opportunities (economic, financial, educational, grants, etc.) in both places opening up for you  

• Experience of living in another place 
• Greater choice of courses 
• Benefits of both the universities 
• Stay with the family during a part of your PhD and its associated benefits (like family support, food, lodging, etc.) 

It provides the best of both worlds. The only thing which stops people from pursuing higher education in India is the common myth that one does not have enough opportunities here. While it might have been true a few decades back, it is the opposite now.

Both the Government of India (GOI) and the private sector have brought many opportunities for Indian graduates in the last decade. For instance, I have been awarded both the Google PhD fellowship (private sector) and the Prime Minister Fellowship (by the GOI). In addition, many companies, including Google, TCS, and IBM, give fellowships to Indian graduate students. But, on the other hand, I have seen my amazing friends in US and UK, who are much smarter than me struggling to bag any such opportunities there.  

There is also a lack of awareness of sponsored PhD programs which allow one to pursue a PhD while working in industry/government. While a decade back, India didn't have many AI labs. This awareness has drastically changed in the last five years. Today almost every big tech company has an AI lab here. Some examples are Adobe, Google, Microsoft, and IBM. 

While earlier grants were limited, now even I have been part of multiple grants to the MIDAS-IIITD lab. For instance, in the last two years, with the help of two grants, I have co-founded two projects: Project Vidhaan for legal language processing and Project Vaani for Indian Spoken Languages. In Project Vidhaan, our core task is to work on legal sublanguage. We have 1 PhD student with four research assistants, three professors, and a few industry collaborators, all spread across the world, collaborating on this project. We started this project with the help of a Technology Innovation Hub (TIH, a GOI grant). In Project Vaani, we want to build open-source Indian spoken language corpora. The government of Singapore, Adobe, the University of Jaffna, and many companies have expressed their commitment to the cause.  

So, today, I don't think there is a lack of opportunities in India. If one has the will, one has a way here. The only thing stopping you is yourself and not something outside. So, opportunities are abundant here, and the time is ripe for innovating and putting something out there.  

I have tried to make the Indian PhD scene a bit clearer in a series of articles on my blog. For instance, I have talked about scholarship opportunities for Indian PhD students and joint PhD programs in India, among other things. 

What advice would you provide to people who wish to pursue a career in artificial intelligence research? What preparations should they make? Can you recommend some books and research papers on AI? 

One should have a solid background in statistics, classical machine learning and the field where you want to apply ML. Therefore, reading classics such as 

• Deep Learning by Goodfellow et al., 
• the first few chapters of PRML by Bishop, 
• Statistical Rethinking by McElreath, and 
• ESL by Hastie is necessary. 

Try the amazing hands-on course by fast.ai to learn to apply neural networks. Then start reading papers from your chosen field, and see how they pick up and frame problems, and after some bootstrap time, you will begin seeing gaps and searching for answers. One thing I would suggest to my past self is to write and communicate more. A few resources I would recommend to learn more about academic writing are 

• The Elements of Style by William Strunk Jr, 
• Stylish Academic Writing by Helen Sword, and 
• Scientific Writing for Computer Science Students by Wilhelmiina Hämäläinen.