Meta
Research Scientist, Artificial Intelligence
$184K–$257K
Menlo Park, CA · Full-time · Posted Sep 1, 2026
- Experience
- 8+ years.
- Education
- Bachelor's degree in Computer Science, Computer Engineering, relevant technical field, or equivalent practical experience; PhD in a related technical field or equivalent depth of research experience is also stated.
About the job
Leads research and engineering on TPU performance and model optimization, translating improvements into production-scale systems for Meta’s first-party models.
What you'll do
- Research and implement TPU optimizations, including Pallas kernels, memory management, and compute efficiency.
- Optimize large-scale distributed training and inference, including MoE and parallelism strategies.
- Set research roadmaps and benchmarking methods, and collaborate on deploying optimizations at scale.
What they ask for
- 8+ years of experience in machine learning systems, model optimization, or high-performance computing research.
- Experience with TPU architecture and optimization, XLA, distributed training, PyTorch, and custom kernels using Pallas or similar frameworks.
- Demonstrated record of transitioning performance research into deployed systems at significant scale.
- Bachelor's degree in Computer Science, Computer Engineering, a relevant technical field, or equivalent practical experience.
Nice to have
- PhD in a related field or equivalent depth of research experience; publication record in relevant systems or AI venues.
- Experience with MoE architectures and production-scale models with billions of parameters.
Pay
$184,000/year to $257,000/year (base salary)
Tools
TPU, Pallas, XLA, PyTorch, Mixture of Experts (MoE)
AI focus
Research and optimize AI model efficiency, training, and inference on TPU, including distributed training and production deployment.
Model-written summary · the original posting is authoritative