Research Scientist Intern, PyTorch Framework Performance (PhD)
**Summary:**
Our team’s mission is to make PyTorch models high-performing, deterministic and stable, via a robust foundational framework that supports the latest hardware, without sacrificing the flexibility and ease of use of PyTorch.We are seeking a PhD Research Intern to work on next-generation Mixture-of-Experts (MoE) systems for PyTorch, focused on substantially improving end-to-end training and inference throughput on modern accelerators (e.g., NVIDIA Hopper and beyond).This internship will explore novel combinations of communication-aware distributed training and kernel- and IO-aware execution optimizations (inspired bySonicMoE and related works) to unlock new performance regimes for large-scale sparse models. The project spans systems research, GPU kernel optimization, and framework optimization, with opportunities for open-source contributions and publication.Team scope:- Improve PyTorch out-of-the-box performance on GPU, CPU, accelerators- Vertical performance optimization for models for training and inference- Model optimization techniques like quantization for improved efficiency- Improve stability and extensibility of the PyTorch frameworkOur internships are twelve (12) to twenty-four (24) weeks long and we have various start dates throughout the year.
**Required Skills:**
Research Scientist Intern, PyTorch Framework Performance (PhD) Responsibilities:
1. Design and evaluate communication-aware, kernel-aware, and quantization-aware MoE execution strategies, combining ideas such as expert placement, routing, batching, scheduling, and precision selection.
2. Develop and optimize GPU kernels and runtime components for MoE workloads, including fused kernels, grouped GEMMs, memory-efficient forward and backward passes.
3. Explore quantization techniques (e.g., MXFP8, FP8) in the context of MoE, balancing accuracy, performance, and hardware efficiency.
4. Build performance models and benchmarks to analyze compute, memory, communication, and quantization overheads across different sparsity regimes.
5. Run experiments on single-node and multi-node GPU systems.
6. Collaborate with the open-source community to gather feedback and iterate on the project.
7. Contribute to PyTorch (Core, Compile, Distributed) within the scope of the project.
8. Improve PyTorch performance in general.
**Minimum Qualifications:**
Minimum Qualifications:
9. Currently has, or is in the process of obtaining, a PhD degree in the field of Computer Science or a related STEM field
10. Deep knowledge of transformer architectures, including attention, feed-forward layers, and Mixture-of-Experts (MoE) models
11. Strong background in ML systems research, with domain knowledge in MoE efficiency, such as routing, expert parallelism, communication overheads, and kernel-level optimizations
12. Hands-on experience writing GPU kernels using CUDA and/or cuteDSL
13. Working knowledge of quantization techniques and their impact on performance and accuracy
14. Must obtain work authorization in the country of employment at the time of hire, and maintain ongoing work authorization during employment
**Preferred Qualifications:**
Preferred Qualifications:
15. Experience working on other ML compiler stack, especially on PT2 stack
16. Familiarity with distributed training and inference, such as data parallelism and collective communication
17. Ability to independently design experiments, analyze complex performance tradeoffs, and clearly communicate technical findings in writing and presentations
18. Intent to return to degree program after the completion of the internship/co-op
19. Proven track record of achieving significant results as demonstrated by grants, fellowships, patents, as well as first-authored publications at leading workshops or conferences such as NeurIPS, MLSys, ASPLOS, PLDI, CGO, PACT, ICML, or similar
20. Experience working and communicating cross functionally in a team environment
**Public Compensation:**
$7,650/month to $12,134/month + benefits
**Industry:** Internet
**Equal Opportunity:**
Meta is proud to be an Equal Employment Opportunity and Affirmative Action employer. We do not discriminate based upon race, religion, color, national origin, sex (including pregnancy, childbirth, or related medical conditions), sexual orientation, gender, gender identity, gender expression, transgender status, sexual stereotypes, age, status as a protected veteran, status as an individual with a disability, or other applicable legally protected characteristics. We also consider qualified applicants with criminal histories, consistent with applicable federal, state and local law. Meta participates in the E-Verify program in certain locations, as required by law. Please note that Meta may leverage artificial intelligence and machine learning technologies in connection with applications for employment.
Meta is committed to providing reasonable accommodations for candidates with disabilities in our recruiting process. If you need any assistance or accommodations due to a disability, please let us know at accommodations-ext@meta.com.