San Francisco Bay Area
My Google Scholar Citation: https://scholar.google.com/citations?user=50UWwt0AAAAJ
Investing in YC startups and VC funds.
Offroad Stack Platform and Integration
IPO'ed in 2020. Led development of computer vision models and pipelines for various DoorDash business cases.
Led Perception team which focused on perception sensor calibration and fusion, object detection and tracking, offline and online perception performance evaluation, deep learning modeling and deployment.
Developed next generation robotics for practical business applications. I was the first autonomy software engineer of DoorDash Labs. - Built the first version of autonomy stack for autonomous delivery robots. The stack included a basic CAN-based controller (which talked directly to driving and steering motor controller), a planner for the robot to travel at a speed of 10 mph, a lidar-based obstacle avoidance system, a GPS-with-heading based localization stack and an interaction engine for robot basic behavior notification. - Built a real-time data streaming pipeline where robot vitals are reported directly to a database. - Structured and implemented KPI metrics (including human intervention, perception false positive rates, etc) to measure robot performance and events for technical and product goals. - Built the first version of deep learning based image detection which works in real-time on the robots.
ACQUIRED BY JOHN DEERE IN 2021: https://www.deere.com/en/our-company/news-and-announcements/news-releases/2021/corporate/2021aug5-bear-flag-robotics/ I involved in self-driving technologies and was the first engineer/employee of Bear Flag Robotics. - Full-stack robotics and perception engineer for BFR autonomous tractor. System-level design for the autonomous tractor including sensor selection (lidar, camera, IMU, GPS, radar, wheel steering encoder) and integration. - Systems integration and algorithm design for localization: Kalman filter for GPS, IMU, radar sensor fusion with odometry readings and visual SLAM benchmarking for localization improvement. - Software architecture and test cases design. - Investigation into utilizing deep learning techniques for object and obstacle recognition.