I make AI compute fast, efficient, and reliable, from the serving layer down to the hardware and the power behind it.
Software Development Engineer II at AWS, New York. At work: fault-tolerant, multi-tenant payment and IoT systems. Outside work: the layer AI runs on.
Merged upstream
- Qwen3-Omni crash on videos with no audio track
Root-caused aStopIterationin multimodal processing whenuse_audio_in_video=True; fixed two follow-on false-positive paths in processor caching and class inheritance. - VMM panic on ACPI device restore
Fixed a panic whenEventFdcreation fails under fd exhaustion. - DNS lookup delay from an IPv6 resolver stall
Documented the cause of the delay. - NIXL metrics: registration-expiry and writer-loop-error counters
Not merged yet.
Selected work
- FlowState
Rust temporal join engine for point-in-time ML feature pipelines, batch and streaming, with zero-copy Arrow. - GPU Flight Recorder
Names the straggling rank in distributed GPU training by joining NCCL collective timings with per-GPU telemetry. Validated in simulation; no real-hardware results yet. - Low-Latency Order Book Engine
Lock-free SPSC ingestion and thread-local slab pools, benchmarked on EC2.