Working in perception production at Uber ATG

Develop numerical method for Vlasov-Maxwell Equations in a HPC setting.

Performed preliminary analysis and experiments of using deep neural networks for data sparse spatiotemporal forecasting.Implemented an efficient data packing algorithm which sped up a frequently used downstream smoothing algorithm by up to 100x.

Developed neural translation system from natural language to software console commands.Began framework for testing different deep learning architectures for a ranking system for answer retrieval in a FAQ.

Experiments in nuclear physics