GPU Kernel Engineer (AI)

Location

Location: San Francisco, Boston, or Remote (Hybrid). While San Francisco and Boston are preferred, other locations are open.

Company

Spun out of MIT CSAIL, building general-purpose AI systems optimized for various deployment targets including data centers and on-device hardware, partnering with enterprises in consumer electronics, automotive, life sciences, and financial services.

What you will do

• Write high-performance GPU kernels for novel model architectures.
• Integrate kernels into PyTorch pipelines including custom ops and benchmarking.
• Profile and optimize training and inference workflows to remove bottlenecks.
• Build correctness tests, numerics checks, and maintain performance benchmarks and guardrails.
• Collaborate closely with researchers to translate ideas into production speedups.

Requirements

• Authored custom CUDA kernels (beyond cuDNN/cuBLAS usage).
• Strong understanding of GPU architecture and performance factors.
• Proficiency with low-level profiling tools like Nsight Systems and Nsight Compute.
• Strong C/C++ programming skills.
• Measurable improvements on critical end-to-end pipelines validated by benchmarks.
• Experience shipping research-driven techniques as production kernels with maintenance.

Nice to have

• CUTLASS experience and tensor core utilization strategies.
• Triton kernel experience and/or PyTorch custom op integration.
• Experience building benchmark harnesses and performance regression tests.

Culture & Benefits

• Unique optimization challenges with high ownership from day one.
• Competitive base salary with equity in a unicorn-stage startup.
• 100% paid medical, dental, and vision premiums for employees and dependents.
• 401(k) matching up to 4% of base pay.
• Unlimited PTO plus company-wide Refill Days throughout the year.