The whole posting in a few lines. Sign up to read it here and on every role you open.
Sign Up to ReadDescription
The company provides GPU virtualization and cloud infrastructure. This role sits on its core engineering team.
The work centers on low-level C++ code that raises GPU utilization for machine learning jobs.
Responsibilities
- Tune the speed of a C++ library that virtualizes GPUs.
- Study GPU oversubscription, checkpointing, and distributed cluster designs.
- Build system-level software that lets the team support new GPU architectures.
- Track down and fix performance problems in production systems across every layer.
- Find where machine learning workloads get held up.
Requirements
- Minimum 3 years of C++ systems engineering. You need deep C++ expertise, or equivalent Rust experience plus a willingness to work in C++.
- Production experience building low-level software such as operating system, kernel, memory, or driver components.
- A proven record of performance tuning in C++ systems or network interface card (NIC) software.
- The ability to trace performance problems across several layers of a system.
- Working knowledge of compilers, networking protocols, kernel internals, distributed systems, infrastructure, or Kubernetes.
- Solid computer science fundamentals and software you have shipped to production, not only academic projects.
Conditions
Base salary of $200,000 to $250,000 per year.
Visa sponsorship is available.
The role is on-site in San Francisco, California.