Research Scientist Intern, FAIR – Multi-Agent Systems and Memory (PhD)

Meta is seeking Research Interns to join our Fundamental AI Research (FAIR) Communication & Language teams working on multi-agent systems and memory. These teams are committed to advancing artificial intelligence through fundamental breakthroughs that enable agents to reason, learn from experience, and collaborate on complex tasks. We are looking for individuals passionate about large language models, agentic systems, machine learning, and reinforcement learning. Our interns have the opportunity to make core algorithmic advances and evaluate their ideas at scale.

This internship will investigate how AI agents can retain and use knowledge across extended interactions, share information effectively, and adapt their workflows based on experience. Our research connects memory and multi-agent collaboration with self-improving systems for automated scientific research. Potential directions include developing compact memory representations, improving inter-agent communication and coordination, and automatically optimizing agent harnesses. The specific research question will be shaped by the intern's interests and early experimental findings.

We are especially interested in candidates with expertise in one or more of the following areas:
Multi-agent systems, collaboration, coordination, communication, and shared knowledge
Agentic memory, long-term memory, implicit or parametric memory, and memory retrieval and utilization
Long-context modeling, context compression and compaction, and KV-cache compression
Self-improving agents, adaptive agent workflows, agent harness optimization, and test-time learning and adaptation
Reasoning, planning, tool use, reinforcement learning, and automated scientific research

Our team at Meta AI offers twelve (12) to twenty-four (24) weeks long internships and we have various start dates throughout the year. To learn more about our research, visit https://ai.meta.com/research/.

Responsibilities

Conduct fundamental research on memory and multi-agent collaboration to improve AI agents on complex, long-running tasks.
Brainstorm with research mentors, review relevant literature and existing systems, and formulate focused research questions.
Develop novel algorithms and architectures for retaining and retrieving context, sharing knowledge across agents, or adapting agent workflows based on experience.
Implement prototypes and conduct rigorous experiments on automated research and related agentic benchmarks, comparing against strong baselines.
Analyze agent trajectories and failure modes to assess memory utilization, coordination, task performance, computational cost, and generalization.
Draft research reports and/or publications and present research outcomes to internal and/or external audiences.
Contribute research that can be applied to Meta product development.

Qualifications

Currently has or is in the process of obtaining a PhD degree in Computer Science, Machine Learning, Artificial Intelligence, Natural Language Processing, or a related field
Research and/or work experience in large language models, machine learning, deep learning, reinforcement learning, or related areas
Experience in Python, C++, 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 Proven track record of achieving significant research results, as demonstrated by grants, fellowships, patents, or publications at leading AI conferences such as NeurIPS, ICML, ICLR, ACL, EMNLP, NAACL, COLM, AAAI, or IJCAI
Research experience in agent memory, context compression, multi-agent systems, agent orchestration, self-improvement, or automated scientific research
Experience building and evaluating language-model agents, including tool use, memory mechanisms, or collaborative workflows
Experience with deep learning frameworks, such as PyTorch, and implementing research prototypes
contributions to open source libraries and frameworks are welcome
Experience designing controlled experiments, analyzing agent trajectories or large-scale experimental results, and using theoretical and empirical research to solve problems
Experience working and communicating cross functionally in a team environment
Intent to return to degree-program after the completion of the internship/co-op