Priya Kansal works at Couger, Tokyo, Japan, as an AI researcher with a non-IT background. She has diverse experience, ranging from academia to industry.

Priya is a self-driven and passionate professional who excels at uncovering hidden gems in significant amounts of structured, semi-structured, and unstructured data. In addition, she has extensive experience in data collection, statistical tool application, interpretation, and visualization techniques.

INDIAai recently spoke with Priya Kansal about her research path, the obstacles she encountered during the learning process, and how she overcame them.

Can you tell us about your diverse educational background, including everything from mathematics to business administration to behavioural analytics?

I started my academic journey as a maths student. However, while becoming a perfect Maths student, I found myself a bookworm who only knew how to solve complex mathematical theorems. This exposure led me to do something different from all this. So I joined a business management program. Since I lived so many years with that analytical mindset, Statistics, Operations Research, and Finance were my favourite subjects. After that, I chose to pursue this interest and opted for MFA. I know that a master's in Business Administration and Financial Analytics can feel like a very different field from AI. Still, while working as an AI researcher, I found that skills like communication, story writing, persuasion and quick absorption of facts are as necessary as the programming skills and all these skills I learned in my masters. Further, I sharpened my analytical talent and acquired skills from my masters during my PhD.

How did you make the transition from academic faculty to AI researcher? What drew you to a career in AI?

During my PhD., the skills I learned during my bachelor's and master's are in full swing. I started enjoying doing the Analysis and writing the reports for myself and fellow researchers. I joined many clubs as a volunteer to work on the research problem of the companies associated with those clubs. But it was all limited to only using statistical software like SPSS, STATA etc. That's when I realized researchers didn't widely use this program in the industry. Then I started learning python programming and R programming. I could know and could use it because I wanted to play with data and drive the stories hidden behind that data. This transition has become my passion now.

What difficulties did you encounter in the early stages of AI research? How did you manage it?

My most significant difficulty was my phobia of programming. I never did this earlier. So, I felt like a nightmare when some implementation parts came to me either on the server or client-side. But, nowadays, resources available online and their accessibility have made it very easy to learn and grow and that too at your desired pace. Moreover, I found that everyone on any platform is very keen to help you if you want to help yourself. 

Can you tell us about some of the exciting challenges you faced while working as an AI researcher?

The most crucial element in AI is data; thus, teaching the data collecting team what information is for business purposes is problematic. Sometimes, maintaining the balance between research and commercial perspectives of the projects is also tricky.

Is a strong programming background necessary to pursue a career in AI? What are your thoughts?

AI is an umbrella term, and in the current market, AI has many roles ranging from AI researcher, AI engineer, Data scientist etc. However, to start with, a strong programming background is not needed, and a person may learn gradually. On the other hand, it's always good to have some programming knowledge /experience. It comes in handy doing small POC before proposing any model/solution, even as an AI Researcher.

What is the one quality that a person must possess to pursue a career in AI?

AI field is relatively new and thus has many real-life use cases to be catered to provide humans with a better life.

Systematic and Analytical thinking is the most important quality a person must possess to pursue a promising career in AI. A curious the box mindset and thirst for problem-solving always result in the best outcomes.

What exactly do you do as an AI researcher at Couger? Could you elaborate on your day-to-day responsibilities at Couger?

Here at Couger, my job is to design the computer-vision related features for our Virtual Human Agent. First, I set the day's schedule, read the research papers, and created the model architecture. Then data requirements, coding, training, testing, iterating the process, if test metrics are not up to the expectation, implementing the model, figuring out the failed cases in production, collecting more data, and then iterating the process again.

Can you share some of the most exciting research articles and books you've come across while researching AI?

I liked the books by Aurélien Géron when I was learning to program. They are pretty intuitive and easy for a beginner, except for this. I wanted Human Compatible AI by Stuart Russel, A Guide for Thinking Humans By Melanie Mitchell. Last year, I started reading articles about biases and fairness in AI as my interest. Currently, I am reading Life 3.0: Being Human in the Age of AI by Max Tengmark.

What advice would you give to students and professionals interested in pursuing a career in AI?

Don't be afraid of trying even if you think you can not do it. Every trail leads you one step nearer to your career goal, even if it fails. Make a list of the job specifications of your dream role and work on acquiring those skills.

As an AI researcher, what is your future goal in AI?

My goal as an AI researcher is to enhance my skills to a level that I can build up solutions for real-life cases that can help humans lead a better life.