Making self driving cars safe through simulation (NPC AI, safety metric estimation, release validation, workflow automation, etc.)

C++, Python, SQL, Beam, data pipelines, data analysis, machine learning

Phishing, spam, and other abuse prevention on Google Drive, Google Groups, Google Photos, Google Docs, and more. GDPR compliance, backend infrastructure, big data analysis, map reduce, C++, Python, Go, SQL, ...

Applied software tools (Python and Jupyter) to automate data collection and statistical analysis, leading to improved efficiency in experimentation for other researchers.

Research work led to 5 papers in peer-reviewed journals, 4 presentations at scientific conferences, and continued funding for two incoming students.

CS231n: Employed Python and TensorFlow to implement convolution neural networks to detect cancer in mammogram images. The project evaluated network architectures for whole image classification and fusing multi view information.

Applied R and C++ to course projects.

Improved runtime of a software profiling tool for massively parallel codes from O(n) to O(logn) allowing users to profile distributed simulations at runtime.

Tested software using ORNL's high performance computing resources.

The code was released with the Community Earth System Model.