Abacus.AI is an end to end MLOps platform that allows users to connect their data, feature engineer, train, serve and monitor their models. Primary expert for custom code in feature engineering, in addition to shipping large improvements in all parts of the product fast.

ONNX Runtime is an open source project that makes it easier to access hardware optimizations. Working on the Execution Provider side to efficiently utilize CPU, GPU, VPU hardware to speed up ML inferencing. Spearheading productive development through end to end pipelines monitoring regressions in latency and memory affecting 200 models. Ongoing contact of an Intel team of contributors in design decisions, usage help, and validation. Co-drove a team of 40+ in product release requiring package generation, validation, and uploading to GitHub, Docker, PyPi, Nuget, and Maven Central. 800K PyPi downloads and 50K Nuget downloads. Skills: ONNX, ONNX Runtime, OpenVINO, TensorRT, CUDA, NVIDIA tools, Docker, Power BI, SQL, Kusto, PyTorch, Netron, Azure

A group of UC Berkeley students that solve problems by building intelligent software. Our mission is to use artificial intelligence and machine learning to develop meaningful technology. Participated on two projects so far. Local Ancestry (in collaboration with Moorjani Lab): Optimized a Hidden Markov Model Algorithm used to describe the ancestry at particular nucleotide (SNP) sites given a genome sequence. Skills: Scikit-learn, Numpy, Matplotlib. Danger Sense: Anticipated dangerous driving situations by combining object detection, an LSTM, and an RL agent. Skills: OpenCV, YoloNet, OpenAI Gym, Keras

Reduced the number of Kusto database queries to investigate incidents from 15+ to 1 by developing an internal tool to track and visualize requests throughout a system. 30+ immediate users, built to be adopted by other teams. Developed an end-to-end service using a C# API calling Kusto queries, with a React frontend to render components. Participated in a weekly Book Club discussing distributed systems internals and design tradeoffs. Skills: C#, Kusto, React.js, CSS, HTML, Injection Dependencies, Azure Storage