AI Compiler

About the role

We are looking for an experienced software engineer to join our AI Compiler team, responsible for building a state-of-the-art MLIR-based compiler stack that unlocks the performance of our next-generation AI hardware.

Key responsibilities

• Design and implement a modern MLIR-based AI compiler targeting proprietary accelerator hardware.
• Define intermediate representations (IRs) in collaboration with platform teams to exploit unique architectural features.
• Work closely with runtime engineers to shape drivers and integration layers for popular AI frameworks.
• Partner with tool-chain groups to deliver end-to-end development workflows, both static and dynamic.
• Contribute to open-source ecosystems and ensure the compiler remains extensible and maintainable.

Required profile

• 4+ years of professional software engineering experience.
• At least 2 years developing back-ends for large-scale, open-source compilers such as LLVM.
• Demonstrated open-source contributions to relevant projects or experience building cutting-edge neural-network architectures.
• Background in ML accelerators, high-performance computing, GPU or CPU architectures.

Required skills

• LLVM compiler infrastructure.
• MLIR backend development (preferred).
• Open-source contribution workflow.
• Neural-network architecture design.
• Experience with ML accelerators, HPC, GPU and CPU architectures.