Kernel Engineer
Recursive Superintelligence · San Francisco · London
Frontier labonsite$200k–$500k
At a glance
Highlights
- Make ML training and inference faster and more efficient at hardware level.
- Uncover bottlenecks across compute, memory, and scheduling to guide optimization.
- Deliver validated improvements with clear before-and-after performance evidence.
What they offer
- Equity
Ideal candidate
Low-level systems thinker who loves reasoning from hardware behavior and owns performance work end to end. You care about rigorous measurement, numerical reliability, and shipping real speedups collaboratively.
Requirements
- 2+ years post-graduation industry experience in GPU/accelerator kernels, ML compilers/backends, or runtime systems
- Deep implementation-level expertise in at least one kernel, compiler, or runtime stack
- Ownership of substantive kernel, compiler, backend, or runtime improvements with measured impact
- Strong understanding of hardware execution fundamentals including memory hierarchy and parallelism
- Ability to reason about numerical correctness, regressions, and performance tradeoffs
- Track record of delivering complex technical work collaboratively from investigation to validation
Lab profile
About Recursive Superintelligence
Building safe, recursively self-improving AI to automate knowledge discovery.
What they build
- AI models
- Tools for automated AI research
Known for
- Recursively self-improving AI
- Open-ended algorithms
- Automated AI research
Posted · Last seen