Meta AR/VR Job | Research Engineer, Pytorch Distributed (PhD)
Job（岗位）: Research Engineer, Pytorch Distributed (PhD)
Type（岗位类型）: Research
Citys（岗位城市）: Menlo Park, CA +1 locations
Date（发布日期）: 2024-10-23
Summary（岗位介绍）
The vision for the PyTorch Distributed team is to make PyTorch the industry leading machine learning framework for Efficient and Large Scale Distributed Training and Inference.
This vision includes:
- Enabling scaling to thousands of GPUs for critical workloads like generative AI and recommendation.
- Work within meta and with leading industry collaborators to ship impact worldwide.
- Outpace our competitors as the preferred framework across the industry for distributed training
- Enable cutting edge research for the largest and most complex models
- Provide easy to use high-level APIs for data, model, pipeline, hybrid parallelism and auto parallelism
- Provide a rich toolkit of basic building blocks that allow exploration of a variety of distributed training and inference paradigms
The team is highly innovative, passionate about the technologies we build, and loves to do deep technical work.
Qualifications（岗位要求）
Currently has, or is in the process of obtaining, a PhD in Computer Science, relevant technical field, or equivalent practical experience.
Proficient in Python, C++ or CUDA programming.
Research or industry experience in ML systems, ML accelerators, HPC, GPU performance, and similar.
Currently has, or is in the process of obtaining a Bachelor's degree in Computer Science, Computer Engineering, relevant technical field, or equivalent practical experience. Degree must be completed prior to joining Meta.
Must obtain work authorization in the country of employment at the time of hire, and maintain ongoing work authorization during employment.
Description（岗位职责）
Develop the Pytorch native parallelism APIs (DDP, FSDP, TP, CP).
Develop large scale infrastructure APIs for collective communication and fault tolerance.
Improve PyTorch performance via systematic solutions for the entire community.
Explore composability of PyTorch distributed with Pytorch compiler and Pytorch core libraries.
Optimize Generative AI models across the stack (pre-training, fine-tuning, and inference).
Conduct cutting-edge research on ML distributed technologies.
Collaborate with users of PyTorch to enable new use cases of PT-Distributed technologies both inside and outside Meta.
Additional Requirements（额外要求）
Familiarity with PT-Distributed technologies or experiences working inside PyTorch.
Expert knowledge in GPU performance and writing high-performance communication libraries and fault tolerance distributed systems.
Proven track record of achieving results as demonstrated by grants, fellowships, patents, as well as first-authored publications at workshops or conferences such as ICML.
