Meta is seeking a Research Engineer to advance the field of artificial intelligence by solving engineering challenges in developing new algorithms and software to accelerate progress towards advanced machine intelligence. The role involves contributing to understanding and improving the societal and economic impact of AI and applying these ideas at scale, particularly at the interface of frontier models and society.
Requirements
- Experience with Python, C or C++ or other related languages and with the PyTorch framework
- Development experience in training large-scale AI models and performance evaluation, and familiarity with model architectures such as Large Language Models, Transformers, and Recommender Systems
- Familiarity with PT2 technologies, Triton, MLIR, or experiences working inside PyTorch
- Expert knowledge in GPU performance and authoring high-performance CUDA kernels
- Experience in solving complex problems and comparing alternative solutions, tradeoffs, and different perspectives to determine a path forward
Responsibilities
- Carry out cutting-edge research to advance the field of machine learning and artificial intelligence
- Contribute research that leads to innovations in: alignment, security and safety, particularly in the context of AI agents
- Devise rigorous engineering and data-driven approaches for efficient training of AI models and algorithms implementation
- Collaborate with research scientists and cross-functional partners in a team environment
- Publish research results and contribute to research that impacts Meta product development
- Open source high quality code and produce reproducible research
Other
- Master's degree in Computer Science, Computer Engineering, a relevant technical field, or 2+ years of equivalent domain-specific industry experience
- Experience working and communicating cross-functionally in a team environment
- Must obtain work authorization in the country of employment at the time of hire, and maintain ongoing work authorization during employment
- First-authored publications at peer-reviewed conferences (e.g., NeurIPS, MLSys, ASPLOS, PLDI, CGO, PACT, ICML, or similar)
