About this role
Lila Sciences is seeking an ML Research Engineer to optimize the performance of large language models (LLMs) on cutting-edge NVIDIA hardware. You will focus on maximizing iteration speed for training and inference runs, ensuring that optimizations are maintainable and accessible to the research team. This role involves deep dives into GPU and cluster performance, big data handling, and establishing performance standards for accelerated scientific discovery.
What we are looking for 6
- Optimize GPU and cluster performance for 1T+ parameter LLMs
- Develop and maintain optimized ML stacks for research teams
- Profile and optimize GPU compute, memory, and collective operations
- Manage and optimize big data handling for large-scale training
- Collaborate on throughput/accuracy tradeoffs for research tasks
- Ensure optimized configurations are tested, accessible, and adaptable