Lead TPU compiler and model optimization for frontier AI inference systems.
Optimize large language models on TPUs across the full execution stack, from high-level model code through low-level kernels to hardware interaction. You'll profile workloads, identify performance bottlenecks, implement kernel improvements, and collaborate with researchers to bring advances into production. Strong background in accelerator-based ML workload optimization required, with expertise in quantization and attention mechanisms highly valued.
Membership is €29/month, cancel anytime: every rate, every original listing link, and a daily alert for roles matching your filters.
Found at a specialist agency · listed 21 October 2026 · InsideJobs links you to the original posting.