Research Intern: Multi-Agent Memory & AI Reasoning

New Grad; Bellevue (WA); Sur place; USD 28 000 - 41 000; Plein temps

Résumé du poste

Meta is seeking Research Interns to advance Fundamental AI Research (FAIR) Communication & Language teams focused on multi-agent systems and memory. Interns will explore how agents retain knowledge across extended interactions, share information effectively, and adapt workflows through experience.

The program spans 12 to 24 weeks with flexible start dates. Ideal candidates have expertise in multi-agent collaboration, memory representations, long-context modeling, self-improving agents,

Qualifications

• Candidates with expertise in multi-agent systems, coordination and communication.
• Expertise in agentic memory, long-term memory and memory retrieval.
• Experience with long-context modeling and KV-cache compression.
• Interest in self-improving agents, test-time learning and adaptation.
• Knowledge of reinforcement learning and automated scientific research.

Responsabilités

• Investigate how AI agents retain and use knowledge across extended interactions.
• Develop compact memory representations and improve inter-agent communication.
• Explore automatic optimization of agent workflows and harnesses.

Connaissances

Multi-agent systems

Agentic memory

Long-context modeling

Self-improving agents

Reasoning and tool use

Description du poste

Meta is seeking Research Interns to advance Fundamental AI Research (FAIR) Communication & Language teams focused on multi-agent systems and memory. Interns will explore how agents retain knowledge across extended interactions, share information effectively, and adapt workflows through experience.

The program spans 12 to 24 weeks with flexible start dates. Ideal candidates have expertise in multi-agent collaboration, memory representations, long-context modeling, self-improving agents,