Modular is on a mission to revolutionize AI infrastructure by rebuilding the AI software stack. The Lead AI Graph Compiler Engineer will extend and develop the Modular graph compiler, setting strategic technical direction and mentoring junior engineers to create a best-in-class multi-device compiler and runtime system.
Responsibilities:
- Set the technical direction for a team building the best heterogenous multi-device compiler and runtime system available, using MLIR and LLVM technologies
- Collaborate with the Mojo compiler team to influence and harness Mojo's powerful compile-time meta-programming capabilities
- Develop and disseminate deep expertise in new hardware platforms, partnering with Mojo kernel developers to achieve state-of-the-art performance on the latest hardware
- Tackle complex technical challenges to empower ML engineers and accelerate model development
- Strategically identify technical opportunities to grow the business
- Mentor and develop junior engineers, growing the capability of the team over time
Requirements:
- 5+ years of compiler engineering experience
- A track record of challenging the status quo and delivering significant, measurable improvements
- Experience working with compilers for machine learning frameworks, such as PyTorch
- Knowledge of core compiler algorithms and data structures
- Knowledge of and experience working with MLIR and LLVM
- Knowledge of C++, as well as, knowledge of basic GitHub workflows like pull requests
- Creativity and curiosity for solving complex problems, a team-oriented attitude that enables you to work well with others, and alignment with our culture
- Experience with ML graph optimizations, parallel / distributed programming, heterogeneous ML computation, and/or code generation
- Advanced degree in Computer Science or a related area is a plus