I've built production infrastructure at Block, DoorDash, and Datadog, including a DAX caching layer for Block's Fidelius tokenization service that reduced read latency by 88% and achieved an 89% cache hit rate. At DoorDash, I created migration observability tooling and daily shard updates for an internal storage system, helping identify overloaded clusters and support more reliable operations.
I've also applied machine learning to biological imaging research at the Broad Institute and NLP at Macmillan Publishers. My work spans Go, Java, Python, AWS, DynamoDB, Terraform, Kubernetes, PyTorch, and TensorFlow.
