Tilde Research
Kernel Engineer (Internship and Full-time)
San Francisco · On-site · Internship · Posted Jul 14, 2025
About the job
Designs and optimizes high-performance GPU kernels and works with ML teams to improve model training and inference performance.
What you'll do
- Design, develop, and tune custom GPU kernels for core model operations.
- Collaborate with ML engineers to prototype and scale novel model architectures.
- Contribute to system-wide improvements in efficiency and throughput.
What they ask for
- Experience in deep learning or related research areas.
- Demonstrated exceptional capability working on ML kernels.
- Deep familiarity with PyTorch and at least one of Triton, TK, or TileLang; basic familiarity with CUDA and knowledge of GPU architecture.
- Clear verbal and written communication, strong algorithmic thinking, and ability to learn quickly.
Nice to have
- Strong open-source contributions, technical blog posts or work logs, or experience with hardware-aligned algorithms.
Tools
PyTorch, Triton, TK, TileLang, CUDA
AI focus
Supports AI research and model development by optimizing GPU kernels for training and inference and helping scale novel model architectures.
Model-written summary · the original posting is authoritative