Meet Amit Kumar, a senior enterprise architect (deep studying) at NVIDIA. Kumar holds a B.Tech in electronics and communication engineering from the distinguished IIT Guwahati. Having labored at among the largest firms like HP and VMWare, he has a wealthy tackle all issues tech.
In an unique interplay with Analytics India Journal, he spoke about his journey in knowledge science, AI and deep studying, whereas taking us over the challenges, achievements and rising developments of this area.
AIM: What attracted you to knowledge science, given that you just have been into software program engineering beforehand?
Amit: I used to be into software program engineering, exactly in pc imaginative and prescient and picture processing area (C++). In 2012, AlexNet competed within the ImageNet Giant Scale Visible Recognition Problem and its success was adopted up by dramatic and speedy development in CNN architectures. This drew me in the direction of deep studying from characteristic engineering-based classical statistical studying. My prior background in picture processing and knowledge principle, and linear algebra helped me get a quicker grip. Equally, the arrival of Word2vec and its cheap success in capturing semantic similarity drew me in the direction of pure language processing.
Transition to the info science area occurred first by classical machine studying, adopted by CNNs (pc imaginative and prescient by deep studying), NLP, Speech recognition(ASR), and eventually reinforcement studying.
AIM: What does your function at NVIDIA entail?
Amit: At NVIDIA, I work as a senior enterprise options architect – deep studying, statistical studying. My main tasks lie in serving to and advising enterprises construct end-to-end knowledge science-based options, ranging from the R&D part (knowledge processing, mannequin coaching) to deployment on NVIDIA AI full stack platforms.
The gamut of enterprises constitutes numerous verticals like healthcare, surveillance, protection, clever video analytics (IVA), good cities, digital twins, AR/VR +AI/ML, industrial visible inspection, good manufacturing, good retail, provide chain logistics, and robotics. Since NVIDIA’s AI platforms are fueled by NVIDIA’s end-to-end knowledge science manufacturing grade and free to make use of AI SDKs and free-to-use fashions, the entire time taken by enterprises to develop and deploy AI options will get drastically lowered and enterprises realise a major acquire relating to ROIs.
For my present function, I usually go to Analytics India Journal to achieve insights into new developments in AI, knowledge science, and the way AI is shaping companies, societies, and policymaking at massive.
AIM: How vital is it for aspirants to start out early or develop their portfolio earlier than venturing into knowledge science and AI?
Amit: A very powerful factor for aspirants is to get the basics proper earlier than diving into knowledge science and AI. Having a fundamental however intuitive understanding of linear algebra, calculus, and knowledge principle helps to get a quicker grip. Aspiring knowledge scientists shouldn’t ignore elementary ideas of software program engineering, on the whole, as a result of these days the market is on the lookout for full-stack knowledge scientists with the aptitude to construct an end-to-end pipeline, reasonably than simply being a knowledge science algorithm skilled.
AIM: What have been among the largest challenges in your profession and the way did you overcome them? Additionally inform us about your skilled achievements?
Amit: My largest problem, which in the end changed into my largest achievement, was to start out from scratch and construct a world-class middle of excellence in knowledge science at HP India together with Niranjan Damera Venkata (distinguished technologist/strategist, AI, and machine studying transformation at HP), Madhusoodhana Rao (director at HP) and Shameed Sait (skilled architect -AI/ML).
This problem was changed into an achievement by going into the start-up mode inside HP. Although we have been half of a big organisation, we made positive that the middle of excellence operates the way in which a profitable startup works by inculcating the tradition of mutual respect and wholesome competitors, attracting and hiring finest skills, and offering freedom and adaptability.
AIM: Everyone desires to be a knowledge scientist. What’s your recommendation to the children beginning out?
Amit: Here’s what I believe:
- Get your fundamentals proper. If that is accomplished, you’re midway by.
- Don’t be a mere ML library person; perceive the algorithm behind it. It will give an intuitive understanding of issues and be of immense assist in devising an answer.
- Do NOT ignore the software program engineering facet of it.
AIM: How do you see the info science and AI house evolving over years?
Amit: Information science and AI house, powered by huge leaps in compute capabilities of GPUs, is simply going to flourish within the coming years. It has already seen its wider addition in numerous segments corresponding to healthcare, good metropolis, retail, governance, protection, schooling, auto-mobile, digital twins, omniverse, AI-powered by simulations, robotics, Business 4.0, and many others.
