Performance Tools Intern
Core
Design and develop a sophisticated performance analysis tool for custom ML accelerators to help engineers understand workload behavior and identify bottlenecks.
Role type
Performance Tools Intern (Hardware/Systems)
Builds
Performance analysis and profiling infrastructure for custom PCIe-based ML accelerators.
Domain
Hardware infrastructure for frontier AI inference
Deliverable
production ML models | product features | dashboards & analysis
Required skills
C++ or Rust programming, computer architecture knowledge, low-level performance analysis, parallel programming, systems software familiarity
Preferred skills
Python, performance analysis tools (Nsight, VTune, Perfetto), ML accelerator architectures, kernel-mode driver development
Technologies
C++, Rust, Python, PCIe, Linux, Nsight, VTune, Perfetto
Responsibilities
Build components of performance analysis infrastructure, collect and analyze performance data from hardware counters and execution traces, develop tooling to trace host-side runtime and accelerator activity, correlate performance events across CPUs and accelerators, build visualization tools for performance bottlenecks.
Seniority
Intern
