Kernelize builds compiler backends for AI accelerators to enable and optimize Triton kernels for broad adoption for ML training and inference. The role involves working on cutting-edge compiler technologies and shaping the core of the technology from the ground up.
Responsibilities:
- Building custom backends for IP accelerators on Triton
- Optimizations
- Code generation
- Pushing performance to the next level
Requirements:
- Passionate about low-level optimizations
- Experience in code generation
- Ability to push performance to the next level
- Experience with compiler technologies
- Ability to work on cutting-edge compiler technologies
- Ability to help shape the core of technology from the ground up