2 months ago
Base Salary
$154k - $217k/yr
Responsibilities
- Own the performance autotuning strategy for MTIA inference and training across current and future chip generations.
- Build and maintain autotuning infrastructure, including search spaces, cost models, benchmarking harnesses, and performance regression detection.
- Define the tradeoff between search cost and performance gains and determine where autotuning belongs in the compile and deployment pipeline.
- Partner with kernel and compiler teams to expose tunable parameters and automate their search.
- Debug and resolve performance issues spanning kernels, runtime, compiler, and serving infrastructure.
- Collaborate with product teams to improve model performance on MTIA.
- Set technical direction and mentor engineers to increase team impact.
Requirements
- Bachelor’s degree in computer science, computer engineering, mathematics, or a relevant technical field, or equivalent practical experience.
- At least 4 years of experience in ML systems, AI infrastructure, performance engineering, or a similar area.
- Strong Python and C++ skills with hands-on experience across the PyTorch stack.
- Experience driving cross-team technical problems where ownership spans multiple groups.
- Knowledge of accelerator performance concepts including roofline analysis, memory bandwidth, occupancy, and kernel performance.
- Track record of setting technical direction and mentoring engineers.
- Preferred experience with torch.compile, TorchInductor, XLA, TVM, MLIR, Triton, or other ML compiler stacks.
- Preferred experience with autotuning, cost models, or search-based optimization such as Ansor, AutoTVM, or learned schedulers.
- Preferred experience with production ML models, hardware bring-up or accelerator development, and kernel-level performance optimization.
- Preferred experience implementing responsible and ethical AI practices and using AI tools, prompt/context engineering, or agent orchestration to improve workflows.
- A master’s or PhD in computer science, computer engineering, compilers/systems, or a related field is preferred.
Benefits
- Bonus, equity, and benefits are provided.
Categories
About Meta
Meta builds social platforms and communication apps—including Facebook, Instagram, WhatsApp, and Messenger—and develops AR/VR hardware and software such as Quest to power immersive computing. It monetizes primarily through advertising tools for businesses, with additional revenue from devices and services, and operates a massive global infrastructure. Founded in 2004 and headquartered in Menlo Park, California, Meta Platforms, Inc. is a public company traded on Nasdaq under the ticker META.