Meet Amit Kumar, a senior enterprise architect (deep learning) at NVIDIA. Kumar holds a B.Tech in electronics and communication engineering from the prestigious IIT Guwahati. Having worked at some of the biggest companies like HP and VMWare, he has a rich take on all things tech. 

In an exclusive interaction with Analytics India Magazine, he spoke about his journey in data science, AI and deep learning, while taking us over the challenges, achievements and emerging trends of this domain.

AIM: What attracted you to data science, given that you were into software engineering previously?

Amit: I was into software engineering, precisely in computer vision and image processing domain (C++). In 2012, AlexNet competed in the ImageNet Large Scale Visual Recognition Challenge and its success was followed up by dramatic and rapid advancement in CNN architectures. This drew me towards deep learning from feature engineering-based classical statistical learning. My prior background in image processing and information theory, and linear algebra helped me get a faster grip. Similarly, the advent of Word2vec and its reasonable success in capturing semantic similarity drew me towards natural language processing.

Transition to the data science field happened first through classical machine learning, followed by CNNs (computer vision through deep learning), NLP, Speech recognition(ASR), and finally reinforcement learning. 

AIM: What does your role at NVIDIA entail?

Amit: At NVIDIA, I work as a senior enterprise solutions architect – deep learning, statistical learning. My primary responsibilities lie in helping and advising enterprises build end-to-end data science-based solutions, starting from the R&D phase (data processing, model training) to deployment on NVIDIA AI full stack platforms. 

The gamut of enterprises constitutes various verticals like healthcare, surveillance, defense, intelligent video analytics (IVA), smart cities, digital twins, AR/VR +AI/ML, industrial visual inspection, smart manufacturing, smart retail, supply chain logistics, and robotics. Since NVIDIA’s AI platforms are fueled by NVIDIA’s end-to-end data science production grade and free to use AI SDKs and free-to-use models, the total time taken by enterprises to develop and deploy AI solutions gets drastically reduced and enterprises realise a significant gain when it comes to ROIs. 

For my current role, I often visit Analytics India Magazine to gain insights into new developments in AI, data science, and how AI is shaping businesses, societies, and policymaking at large.   

AIM: How important is it for aspirants to start early or develop their portfolio before venturing into data science and AI?

Amit: The most important thing for aspirants is to get the fundamentals right before diving into data science and AI. Having a basic but intuitive understanding of linear algebra, calculus, and information theory helps to get a faster grip. Aspiring data scientists should not ignore fundamental principles of software engineering, in general, because nowadays the market is looking for full-stack data scientists with the capability to build an end-to-end pipeline, rather than just being a data science algorithm expert.

AIM: What were some of the biggest challenges in your career and how did you overcome them? Also tell us about your professional achievements?

Amit: My biggest challenge, which ultimately turned into my biggest achievement, was to start from scratch and build a world-class center of excellence in data science at HP India along with Niranjan Damera Venkata (distinguished technologist/strategist, AI, and machine learning transformation at HP), Madhusoodhana Rao (director at HP) and Shameed Sait (expert architect -AI/ML). 

This challenge was turned into an achievement by going into the start-up mode within HP. Though we were part of a large organisation, we made sure that the center of excellence operates the way a successful startup works by inculcating the culture of mutual respect and healthy competition, attracting and hiring best talents, and providing freedom and flexibility.

AIM:  Everybody wants to be a data scientist. What’s your advice to the youngsters starting out?

Amit: Here is what I think:

• Get your basics right. If this is done, you are halfway through.

• Do not be a mere ML library user; understand the algorithm behind it. This will give an intuitive understanding of problems and be of immense help in devising a solution.
• Do NOT ignore the software engineering aspect of it.

AIM: How do you see the data science and AI space evolving over years?
Amit: Data science and AI space, powered by enormous leaps in compute capabilities of GPUs, is only going to flourish in the coming years. It has already seen its wider addition in various segments such as healthcare, smart city, retail, governance, defense, education, auto-mobile, digital twins, omniverse, AI-powered by simulations, robotics, Industry 4.0, etc.