RESEARCH

Meta Superintelligence Labs is seeking a Research Scientist Manager to join our team and work on the next generation of large language models, particularly focusing on MetaAI Assistant Core Capabilities Post-Training. As a Research Scientist Manager, you will play a critical role in building our series of Llama models empowering MetaAI Assistant with a focus on Post-training to shape the raw broad intelligence of pre-trained models into high performance on the most useful skills. It requires dedicated research and development work on specific critical capabilities, as well as creation and validation of new fine-tuning and reinforcement learning methods to enable best models

MINIMUM QUALIFICATIONS

• 12+ years of, or PhD + 8 years of hands-on experience in large language model, NLP, and Transformer modeling, in the setting of both research and engineering development
• Experience and track record of landing large research and/or product impacts in a time-sensitive environment
• 2+ years of managing managers, 5+ years of hands-on supporting and leading teams of research scientists and software engineers
• Proven technical vision in where the field of generative AI will go
• Experience of and knowledge of data curation techniques (training set preparation, ablation experiments, etc.)
• Experience with cross functional collaboration with other teams including non-engineering functions
• Demonstrated experience recruiting, building, structuring, leading technical organizations, including performance management

PREFERRED QUALIFICATIONS

• PhD in deep learning, artificial intelligence, and/or related technical field
• Experience and knowledge of ML frameworks like PyTorch, TensorFlow, etc
• Experience and knowledge of large-scale data platforms such as Spark, Hive, etc
• Experience and knowledge of working with LLM frameworks like LangChain
• Experience and knowledge of training LLMs, fine-tuning on datasets, especially Llama
• Experience with cross functional collaboration with other teams including non-technical functions
• Demonstrated experience recruiting, building, structuring, leading technical organizations, including performance management
  For those who live in or expect to work from California if hired for this position.

• Drive end-to-end development of large language models, including data sourcing, experiment design, evaluation and more
• Drive efficiency gains on training and deployment of LLMs through novel techniques
• Lead a team of applied researchers to democratize Llama for Meta’s users
• Communicate, collaborate, and build relationships with clients and peer teams to facilitate cross-functional projects
• Remain up-to-date on ongoing research and software development activities in the team, help work through technical challenges, and be involved in design decisions
• Remain involved in the research community, both understanding trends, and setting them