Research Scientist Intern, MSL - Agentic Post-Training (PhD)

Meta was built to help people connect and share, and over the last decade our tools have played a critical part in changing how people around the world communicate with one another. With over a billion people using the service and more than fifty offices around the globe, a career at Meta offers countless ways to make an impact in a fast growing organization. Meta is seeking Research Scientist Interns to join a post-training team in Meta Superintelligence Labs (MSL). The team focuses on agentic post-training of Muse foundation models, as part of our mission to build personal superintelligence. Our team’s work covers end-to-end model post-training cycles, including supervised fine-tuning (SFT), reinforcement learning (RL), evaluation, and agentic task and environment creation. We are looking for candidates who are passionate about improving the model's agentic capabilities, including but not limited to long-horizon task completion, complex tool use, search and deep research, and browser use. As an intern, you will have the opportunity to make core model improvements, apply your ideas at unprecedented scale, and directly shape the user experience of Muse. Our internships are twelve (12) to twenty-four (24) weeks long, and we have various start dates throughout the year. Responsibilities Develop novel, state-of-the-art agentic post-training methods and supporting systems, using deep learning and reinforcement learning techniques. Work hands-on across the full LLM post-training stack, and execute, analyze, and iterate on large-scale post-training runs. Analyze and improve the efficiency, scalability, and stability of agentic training algorithms and deployed systems. Advance the science and technology of intelligent agents that can reason, use tools reliably, and complete long-horizon tasks. Collaborate with researchers and cross-functional partners, including communicating research plans, progress, and results. Share research results through publications, presentations, open-source releases, or model contributions. When applicable, contribute research that can be applied to Meta product development. Minimum Qualifications Currently has, or is in the process of obtaining, a Ph.D. degree in Computer Science, Artificial Intelligence, Machine Learning, Natural Language Processing, or a relevant technical field Research experience in machine learning, deep learning, or natural language processing Experience with deep learning frameworks such as PyTorch Experience with Python, C++, C, Java, or other related languages Must obtain work authorization in the country of employment at the time of hire and maintain ongoing work authorization during employment Preferred Qualifications Intent to return to the degree program after the completion of the internship/co-op Proven track record of achieving significant results, as demonstrated by grants, fellowships, patents, first-authored publications at leading conferences, or previous internship projects Research experience or publications in large language models, foundation models, post-training, agentic AI, or reasoning/coding Experience with reinforcement learning and reward modeling Experience with agentic systems, such as agent workflows, tool use, and RL environments, including core contributions to open-source libraries and frameworks for agentic LLMs or RL environments Experience with LLM evaluation, including setting up ML experiments and analyzing their results Demonstrated engineering experience through an internship, work experience, coding competitions, or widely used contributions to open-source repositories (e.g., GitHub) Experience working and communicating cross-functionally in a team environment $7,650/month to $12,667/month + benefits