Having worked across the entire stack of Autopilot and Optimus, my most notable contributions are:

* Real world RL experiments and training pipelines.

* Complete redesign of telemetry feature flags for the tesla fleet.

* Embedded modern C++ for real-time tele-operation, linux system stability, and network security.

* Petabyte level backend infrastructure for fleet data ingestion, processing, querying, and storage.

C++ SDK for multithreaded low-latency deep learning inference server on Xavier SoC. Drive IX/Jarvis project.

Active size-constrained graphical model learning algorithms

Publications:

Robust Plant Phenotyping via Model-based Optimization (IROS 2018).

High-throughput Robotic Phenotyping of Energy Sorghum Crops (FSR 2017).

3D plant point cloud simulation in Matlab. Feature extraction in Tensorflow/Torch.

Browser compatibility testing in Selenium Webdriver.

Anomaly detection with Azure ML, Trident Engine, Bing Dynamic Crawler.