I am working at Annapurna Labs on inference optimization for AWS Neuron, the SDK powering custom AI accelerators at Amazon Web Services. My recent work spans disaggregated inference, speculative decoding, sharding techniques, and megakernels.

I interned at Kensho Labs, ZKH, United Imaging Intelligence, and EnterviewAI. My work there covered core infra, ML infra, backend systems, and agentic systems.