GPU Kernel Engineer

About the role

TypeSafe AI is building reliable, general AI systems that power economically valuable automation. As a GPU Kernel Engineer you will accelerate both training and inference of frontier models by writing and optimizing low‑level CUDA kernels.

Key responsibilities

• Write, optimize, and maintain high‑performance GPU kernels (e.g., CUDA, CuTe DSL) for model training and inference.
• Profile end‑to‑end performance, identify bottlenecks, and implement fixes across the stack.
• Collaborate closely with research and platform engineers to maximize throughput and minimize latency on our hardware.

Required profile

• Deep expertise in CUDA and GPU kernel development with a proven record of performance improvements.
• Experience building and optimizing inference and training kernels for large language models.
• Hands‑on LLM training experience at production scale.
• Ability to reason from first principles about performance, memory, and parallelism.
• Responsible, ownership‑oriented team player aligned with the mission.

Required skills

• CUDA
• CuTe DSL
• GPU kernel development
• Performance profiling
• LLM training experience

What we offer

• Base salary $150k–$250k plus equity, based on leveling.
• 100% covered health insurance.
• Daily lunch and dinner.
• Visa sponsorships.
• 401(k) plans.