I’m Arunabh, a data scientist at Arista Networks, where I build AI systems that transform how engineers work with complex technical information.
My focus is on bridging the gap between cutting-edge research and production-ready systems. I develop LLMs that are accurate enough for mission-critical tasks, fast enough for demanding engineers, and robust enough to handle the messy reality of technical documentation.
I earned my MS from UCLA after my undergraduate in Electrical Engineering at IIT Bombay. My background also includes research at EPFL in Switzerland and building a global learning platform for Google Summer of Code that now serves thousands of students.
My experience spans from reconstructing protein structures to developing production AI systems. While this work has led to patents and publications, I’m most rewarded by building tools that people use to solve real-world problems.
When I’m not building AI, I enjoy experimenting with new technologies and writing to make complex topics more accessible. You can also often find me in the kitchen, trying out new recipes.
Posts
-
Pathways: Branching AI Conversations
An experimental project exploring branching, non-linear conversations with AI: enabling parallel exploration without losing context. -
Instant Pot Pav Bhaji
A delicious Pav Bhaji recipe made in the Instant Pot. This popular Mumbai street food features a spiced vegetable mash served with butter-toasted buns. -
Multi-Agent AI Systems Tackle Personal Finance
An experiment using specialized AI agents, orchestrated to work together and provide more comprehensive insights than a single Large Language Model (LLM) might provide alone. -
Network Endpoint Entity Resolution Patent
A patented system that automatically resolves device identity ambiguity at scale, enabling precise security policies without manual configuration. -
Ab Initio Protein Structure Reconstruction
A novel algorithm that automatically reconstructs multiple protein conformations from mixed 2D projections without knowing viewing angles or the number of structures present. -
Building the Oppia Learner Dashboard
A foundational project for the Oppia Foundation during Google Summer of Code 2017 to build a centralized dashboard, enhancing user retention and creating a personalized learning journey.
subscribe via RSS