Software Engineer, Performance Instrumentation
Core
Define and maintain performance data requirements, identify visibility gaps, and build low-overhead instrumentation and annotation APIs for profiling libraries, frameworks, and runtimes.
Role type
Senior IC software engineer (performance instrumentation)
Builds
Low-overhead instrumentation and annotation APIs for AI and accelerated computing workloads
Domain
AI, accelerated computing, GPU systems
Deliverable
production ML models
Required skills
CUDA, Python, C++, machine learning frameworks (PyTorch, JAX), annotation/tracing APIs (NVTX), data pipeline validation, instrumentation overhead measurement, cross-component debugging
Preferred skills
Telemetry conventions/schemas, storage/checkpointing performance, low-overhead event collection, network/collective-communication telemetry (NCCL)
Responsibilities
Define performance data requirements from various sources; Identify and prove visibility gaps in current data; Inform the build of low-overhead instrumentation APIs; Ensure data quality at ingestion via validation and alerting; Measure and manage instrumentation overhead on real workloads; Collaborate with partner teams to manage commitments and progress
Seniority
Senior, hands-on IC
