3 days ago
Menlo Park, CA, USAMid Level
H1B sponsor
Base Salary
$122k - $181k/yr
Responsibilities
- Improve PyTorch training, post-training, and inference performance on modern AI hardware accelerators.
- Enable new models in TorchTitan for pre-training and post-training.
- Optimize TorchTitan pre-training and post-training for current generative AI models.
- Develop parallelism APIs, kernels, and tools for debugging and visualizing large distributed training jobs.
- Build distributed training infrastructure for communication collectives, profiling, and memory allocation and management.
Requirements
- Currently hold or be pursuing a Bachelor's degree in Computer Science, Computer Engineering, or a relevant technical field, or have equivalent practical experience; the degree must be completed before joining Meta.
- Demonstrate proven Python programming ability.
- Have experience developing AI frameworks or accelerating deep learning models on hardware architectures.
- Preferred: experience implementing responsible and ethical AI practices, including risk assessment, bias mitigation, and quality and accuracy reviews.
- Preferred: knowledge of GPU, CPU, or AI hardware accelerator architectures.
- Preferred: experience with PyTorch, Caffe2, TensorFlow, ONNX, or TensorRT.
- Preferred: AI high-performance kernel experience with CUDA, OpenMP, OpenCL, or AI hardware accelerator kernel programming, including libraries such as cuBLAS, cuDNN, CUTLASS, HIP, or ROCm.
- Preferred: experience developing AI training and inference framework components and analyzing latency, memory bandwidth, I/O access, compute utilization, and related tooling.
- Preferred: demonstrated ongoing AI skill development, workflow optimization using AI tools, and measurable impact on efficiency or quality.
Benefits
- Base salary of $121,992/year to $181,000/year, plus bonus, equity, and benefits.
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.
