Summary
Anurag Paul is a Sr. Staff Research Engineer specializing in multi-modal sensor fusion and real-time perception for autonomous vehicles, with a decade of experience translating ML research into production-grade systems. He has led architecture and optimization of BEV transformers, implemented INT8/FP16 edge inference pipelines, and scaled training workflows to leverage 600K+ unlabeled images and synthetic data for rare-case robustness. His work reduced lane-keeping failure modes in severe occlusion, eliminated yaw-flipping through a novel two-bin-yaw formulation, and accelerated data preprocessing from an hour to four minutes via Lance and Ray. As a technical leader he enforces strict engineering practices (CI/CD, near-100% test coverage), manages cross-functional ML teams, and is an inventor on eight US patents in AV perception and trajectory prediction. Based in California, he bridges academic rigor (MS in ML & Data Science from UCSD) with operational impact, routinely presenting PlusAI’s perception stack at industry conferences.
10 years of coding experience
13 years of employment as a software developer
University of California, San Diego
Indian Institute of Technology Roorkee
Hindi, English