About me

I am a Staff Machine Learning Researcher at SandboxAQ. The overall goal is to develop Large Quantitative Models that are based on AI & Quantum simulations. My focus is on developing graph based LQM technology for efficient drug discovery.

I worked as a Scientist at Microsoft Research (2021-2024). I have been a core contributor to MSR's Project Science Engine. My work was on advancing Generative AI, Graph Neural Networks for Genomics and designing efficient Large Language Models for enterprise using high performance computing techniques.

I was fortunate to be the first PhD in Machine Learning graduate from Georgia Tech (2016-2020). My advisors were Prof. Srinivas Aluru and Prof. Le Song.  My thesis was titled "On using inductive biases for designing deep learning architectures".

In the past, I have been a scientist at Xerox Research (2014-2016). I developed methods to solve the class imbalance problem in ML and designed AI devices for ICU patient monitoring.

My undergrad degree was in Electronics & ECE from Indian Institute of Technology Kharagpur (2009-2014). I enjoy swimming and racquet sports like squash, tennis & badminton.

Email: contact [at] harshshrivastava [dot] com