Lemurian Labs is reimagining the foundations of computing to make AI accessible to everyone. They are seeking a Graph Optimization Compiler Engineer to design and implement optimization passes in their AI compiler stack, focusing on graph-level transformations to enhance model performance.
Responsibilities:
- Design, develop, and maintain the graph optimization layer of our heterogeneous AI compiler
- Implement and extend graph-level transformation passes including operator fusion, layout propagation, dead code elimination, constant folding, and algebraic simplification
- Define and evolve our intermediate representation (IR) to support new optimization opportunities as ML model architectures advance
- Analyze performance data to identify optimization gaps and drive measurable improvements in throughput and latency
- Collaborate with front end and code generation teams to ensure clean IR interfaces and well-structured optimization pipelines
- Propose and prototype new optimization strategies in response to advances in model design and hardware capabilities
- Contribute to testing and validation infrastructure to ensure optimization correctness across model types and hardware targets