John brings experience in performance and efficiency, from highly distributed systems to low level deep learning operators.
At Facebook, John worked on the distributed tracing platform, building systems to collect and process vast amounts of trace data. He focused on increasing observability of long running streaming services like message queues and ML pipelines. Most recently, he worked on PyTorch performance for on-device inference, enabling a range of computer vision use cases.
John holds an A.B. in Computer Science and Statistics from Harvard.