Meta is forming an elite AI research unit, MRS Research, within the Meta Recommendation Systems group to optimize Facebook and Instagram recommendation algorithms and content personalization.

MRS Research will pursue long-term AI research goals and publish cutting‑edge work, closely collaborating with the Ads division to boost Meta’s recommendation and ad systems.

The creation of MRS Research follows a October 2025 reorganization of Meta’s MRS, signaling a strategic push to leapfrog current AI-driven recommendations.

The team is attracting top talent from major tech players, including Lihong Li from Amazon, Xiaolong Wang from OpenAI, Fei Sha from Google, and hires from Thinking Machines Lab.

Meta provided a standard statement from a spokesperson; individual prospects did not respond to inquiries.

There are no specific dates mentioned beyond the article context, but recruitment and development are ongoing.

MRS Research’s mandate covers content understanding, user understanding, retrieval, and ranking across Facebook, Instagram, Threads, and other Meta platforms.

Meta’s three-year plan contemplates spending over $600 billion in the U.S. on AI through 2028, including the acquisition of a 49% stake in Scale AI and appointing Alexandr Wang to head Meta Superintelligence Labs.

Yang Song, previously at TikTok, leads MRS Research as vice president of recommendation research, joining Meta in late 2025.

Talent for MRS Research is already drawing from TikTok and Amazon, signaling a significant influx of experienced researchers.

Meta has been ramping up AI talent since mid-2025 with hires from Thinking Machines Lab and ongoing internal AI branding and training to integrate AI across products.

Concurrent with MRS AI formation, Meta hired Andrew Tulloch from Thinking Machines Lab and added two more AI researchers.