Oakland, California, United States
I’m passionate about building ML infrastructure that makes AI workloads fast, reliable, and easy to use—because AI is transforming productivity and accelerating progress. At NERSC, I build AI inference systems across a supercomputer, improving serving scalability, performance, and observability. My work spans deployment tooling, monitoring and debugging, and partnering with platform teams and users to make complex workflows reliable and high-performing.
Building and scaling AI inference infrastructure on HPC systems at NERSC, while partnering with researchers and developers to deliver reliable, high-performing workflows.
Developed and streamlined Jupyter-based analysis workflows for dark matter detector data, improving usability and maintainability while mentoring a PhD student.
Built and maintained scientific software and data pipelines in a large collaboration, spanning development, testing, documentation, and user enablement.
Automated Python workflows for data collection and analysis, improving reliability and reducing manual effort.
Built and evaluated a CNN-based particle reconstruction workflow and benchmarked end-to-end performance on NERSC systems.