Summary
Siddharth Agarwal is a perception engineer with 9 years of hands-on experience building real-time sensor-fusion and perception systems for autonomous vehicles, currently leading Perception SW at XPENG in the San Francisco Bay Area. He has deep C++ expertise integrating cameras, radar, stereo, IMU and ultrasonics to produce robust local object maps, and previously drove ADAS and autonomous-driving efforts at Canoo. Trained in ECE at Georgia Tech, his background blends academic research in adversarial robustness and control with practical ML work—training YOLO models, accelerating detection to 45 FPS, and developing auto-annotation pipelines using Mask R-CNN. He has a track record of improving system robustness through Kalman-filter-based multi-object tracking, data association, and ensemble anomaly-detection techniques. Colleagues describe him as a pragmatist who bridges research and production: he pairs algorithmic rigor with engineering discipline to ship low-latency perception stacks. An interesting side note: early research on adversarial boosting boosted MNIST adversarial accuracy from 67.5% to 80%, highlighting his focus on resilient ML systems.
9 years of coding experience
4 years of employment as a software developer
Bachelors in Engineering, Instrumentation and Control Engineering, 77%, Bachelors in Engineering, Instrumentation and Control Engineering, 77% at Netaji Subhas Institute of Technology
Masters in Electrical and Computer Engineering, Computer Vision, Machine Learning and Robotics, Masters in Electrical and Computer Engineering, Computer Vision, Machine Learning and Robotics at Georgia Institute of Technology
Maharaja Agrasen Model School
English, German, Hindi