Research Scientist Intern: Agentic AI & LLMs

Meta; Paris; Sur place; EUR 11 000 - 22 000; Plein temps

Résumé du poste

Meta is seeking Research Interns to join the Meta Superintelligence Lab in one of the post-training modeling teams, focusing on Agentic AI, Muse, and personal superintelligence. Internships last 12 to 24 weeks with several start dates yearly.

The role offers the chance to develop state-of-the-art agentic AI algorithms, prototype features at scale, and collaborate with researchers across cross-functional teams. A strong publication record is valued.

Qualifications

• Masters degree (or in progress) in CS/AI/generative AI or related field.
• Experience with Python.
• Experience building systems with ML/DL methods.
• Proven track record of impactful research (grants, fellowships, first-authored papers).
• Experience communicating across teams and with partners.

Responsabilités

• Develop novel agentic AI algorithms and systems using ML/RL techniques.
• Research agentic LLMs, RL environments, post-training and related topics.
• Improve efficiency, scalability, and stability of agentic AI algorithms and systems.
• Advance agentic AI towards reasoning, tool use, and personalized interactions.
• Collaborate with researchers and cross-functional partners; communicate plans and results.
• Disseminate research through publications, presentations, and open source.
• Contribute to research applicable to Meta product development.

Connaissances

Python

Deep learning frameworks

Cross-functional teamwork

Communication of research

Agentic AI research

Publications/Patents/Conferences

Formation

Masters in Computer Science / Artificial Intelligence / Generative AI

Description du poste

Meta is seeking Research Interns to join the Meta Superintelligence Lab in one of the post-training modeling teams, focusing on Agentic AI, Muse, and personal superintelligence. Internships last 12 to 24 weeks with several start dates yearly.

The role offers the chance to develop state-of-the-art agentic AI algorithms, prototype features at scale, and collaborate with researchers across cross-functional teams. A strong publication record is valued.