We are seeking a Research Scientist at the Staff level (IC6) with deep expertise in TPU performance optimization, large-scale model training, and systems-level machine learning. In this role, you will lead high-impact research on model efficiency and optimization for first party models within Meta's native PyTorch stack, collaborating across research and engineering teams to drive AI capabilities that define Meta's next generation of products and platforms.
Responsibilities
Lead the design and execution of TPU performance optimization research, including kernel development, memory optimization, and compute efficiency improvements
• Develop and optimize Pallas kernels for large-scale model training and inference on TPU architectures
• Drive model optimization techniques including Mixture of Experts (MoE), tensor parallelism, pipeline parallelism, and other distributed training strategies
• Optimize first party models within Meta's native PyTorch stack, ensuring efficient integration with XLA compilation and TPU execution
• Identify and resolve complex technical challenges in model training efficiency, inference latency, and system reliability that require novel approaches
• Define and drive multi-quarter research roadmaps for TPU optimization, aligning project milestones with broader organizational goals
• Establish rigorous experimentation frameworks for performance benchmarking, including metric selection, profiling methodology, and data-driven optimization decisions
• Translate research findings into production-ready optimizations by collaborating with engineering teams on deployment pipelines and reliability at scale
• Communicate research findings and technical trade-offs clearly through publications, design documents, and presentations to both technical and non-technical audiences
• Mentor other researchers and engineers on TPU optimization techniques, providing structured feedback on technical direction and experimental rigor
Minimum Qualifications
• Bachelor's degree in Computer Science, Computer Engineering, relevant technical field, or equivalent practical experience
• 8+ years of experience in machine learning systems, model optimization, or high-performance computing research
• Experience with TPU architecture and performance optimization, including profiling, kernel development, and memory management
• Experience with XLA compilation, graph optimization, and low-level performance tuning for accelerator hardware
• Experience developing and optimizing large-scale distributed training systems, including parallelism strategies such as data, tensor, and pipeline parallelism
• Experience with PyTorch and its integration with accelerator backends
• Experience communicating complex technical findings in writing, including technical reports, design documents, or peer-reviewed publications
Preferred Qualifications
• Experience developing custom kernels using Pallas or similar kernel authoring frameworks for TPU or GPU
• Demonstrated track record of transitioning performance research into deployed systems used at significant scale
• PhD in Computer Science, Machine Learning, Computer Architecture, or a related technical field, or equivalent depth of research experience
• Publication record in systems for ML venues such as MLSys, OSDI, SOSP, or related AI conferences such as NeurIPS, ICML, or ICLR
• Experience with Mixture of Experts (MoE) architectures and their optimization for efficient training and inference
• Experience optimizing production-scale models with billions of parameters
