My tryst with Artificial Intelligence began in 1996-97 while working on Offline Cursive Handwriting Recognition at the Indian Institute of Technology, New Delhi. Later, in graduate school, I worked on a project to enable better communication for children with cerebral palsy with Dr. Richard Foulds and Dr. Daniel Chester, at Alfred I. Dupont Hospital for Children. We mapped children’s gestures, expressions, and speech into clear messages for assistive robots at the Applied Science & Engineering Labs. I then had a brilliant experience helping Dr. Keith Decker teach a graduate-level course on AI for four semesters in a row!
I got introduced to data science when Dr. Dan Chester encouraged me to work on the analysis of clinical trials data of an anti-HIV drug at DuPont Pharma’s Stine Haskell Research Center. I then delved into dynamic mathematical modeling with partial delay differential equations, statistical design of experiments and machine learning while completing my Ph.D. at University of Delaware and in post-doctoral research at CalTech.
When I joined Intel Corporation, I was asked to create design rules for reliability for a new line of processors finding risks using large Monte Carlo models. I was flummoxed because despite having a Ph.D., I had no prior experience in this field and needed help even with simple Microsoft Excel functions like VLOOKUP. Although initially reluctant, I approached colleagues and my manager and learned things from scratch. It was a humbling experience, and since then, I have felt more comfortable asking for help when I need it.
On joining Oracle in 2015, I quickly learnt SQL and had the opportunity to write six thousand lines of SQL to solve a pressing problem in three weeks. We were struggling with the calculation of the p-value for Fisher’s Exact test for statistical significance, and I went so far as to refer to Srinivas Ramanujan’s lost diary to find his approximation for the Gamma function.
I enjoy working with Oracle as it gives me a winning edge – the privilege of learning and working with smart creative colleagues and supportive leaders who are committed to solving problems that give directly actionable insights. Independent of designations, I aspire to continue being a facilitator, problem solver, and change maker.
I believe that communication is of utmost importance, whether it’s between stakeholders or internal teams – especially for data driven decision making. When it comes to professional learning, I make it a point to stay immersed in new offerings and knowledge, either online or offline, while honing existing skills. There is no better way to grow than to get your hands dirty with an end-to-end project and understand the travails of real-world work.
Among new technology that fascinates me, I’m awed by Microsoft’s recent breakthrough in General Language Understanding Evaluation using Multi-Task Deep Neural Network (MT-DNN) ensemble model and am keen to learn more.
I am a quiet person who loves to listen and understand as I have always believed that it is better to remain silent and be thought a fool than to speak and remove all doubt. A family man through and through, I love spending time with my wife and kids. I would someday like to transform how we teach emotional intelligence to children.
Due to the bug of curiosity I have ended up working in multiple fields from Electrical Engineering to Computer Science, Computational Systems Biology to a mix of Mechanics, Semiconductor Physics, Chemistry, Electronics, and Statistics, to Artificial Intelligence and Data Science. So, I end up asking a lot of fundamental or even basic questions for I really don’t know. Sometimes, that is exactly what we need to define and formulate problems better.
Vikas Agrawal is a featured speaker at DataHack Summit 2019 – India’s Largest Applied AI & ML Conference happening from 13th to 16th November 2019 in Bengaluru.