Justin Yang

I’m a Senior infrastructure engineer with 10+ years of experience building reliable platforms for large-scale production systems. My work spans container observability and tracing, ML platform development, and GPU cluster infrastructure—always grounded in scalable design and hands-on technical ownership.

At NVIDIA, I built core AI infrastructure across DGX SuperPOD and Base Command Manager, focused on provisioning, workload orchestration, telemetry, reliability, and multitenant GPU platform operations. I automated node bring-up and service registration, translated cluster state into scheduler-ready resources for Kubernetes and Slurm, and implemented health monitoring around GPU, node, and fabric signals so operators could respond faster.

Before that, at Databricks, I developed ML platform capabilities across Feature Store and MLflow Model Registry. I built feature-table and training-set workflows, implemented model packaging with feature lookup metadata, and connected Delta/Spark pipelines to low-latency serving stores for offline-to-online publishing.

Earlier at Datadog, I started in reliability engineering and supported container monitoring and early APM, with ownership across agent operations, ingestion reliability, deployment automation, and production response. I still bring that operational mindset—designing telemetry, automation, and runbooks so distributed systems stay dependable under real-world load.