This list contains only the countries for which job offers have been published in the selected language (e.g., in the French version, only job offers written in French are displayed, and in the English version, only those in English).
Meta is seeking hands-on engineering managers to join the Meta SuperIntelligence Lab (MSL), making direct contributions to the next generation of Generative AI models. The MSL Infra Optimizations team’s mission is to enable the development and productionization of cutting-edge high-performance optimizers, kernels and numeric algorithms, providing comprehensive analyses of their effect for MSL. Join us and be a part of the team that is shaping the future of Meta Superintelligence Lab’s infrastructure!
Job Responsibility:
Lead and support the team that develops various kernels including but not limited to GEMMs, Attention mechanisms etc. Also, contribute to enabling performance at scale of our inference and training of next generation GenAI (Llama) models
Enable the growth of individual contributors, driving the technical roadmap along with technical leads and expand the impact of the team by growing new skill-sets and capabilities
Lead a high performance team of engineers to deliver new capabilities and efficient compute systems for our fleet
Technical management
Experience in systems architecture, performance, workload-analysis and large scale distributed systems
Work cross-functionally across hardware and software/services team to drive engineering efforts
Requirements:
MS or BS in Computer Science or Electrical/Electronics Engineering or equivalent
3+ years of experience of directly managing or leading a team of engineers with varied skill levels
Experience in leading teams working on high performance computing (HPC) and AI/ML systems, including: GPU/ASIC-based kernel development and optimization (e.g. CUDA)
Distributed systems for large scale training and serving
Systems Architecture + Performance
Large scale distributed systems
Experience running a large-scale program and dealing with ambiguity
Nice to have:
Familiarity with the latest techniques in optimizing GenAI workloads
Using frameworks like PyTorch, TorchTriton to develop custom kernels
Understands Kernel enablement and optimizations, including experience working on attention kernels
Understanding GPU memory hierarchy and computation capabilities
Understands low-level CUDA kernel optimizations for inference and training
Experience with Quantization and structure sparsity for low precision training & inference
Understands Optimizers such as Adam, Shampoo, Muon