Tuning Your Algorithm with MRS Memory System and Think-Then-Recommend (TTR) | Qi Guo from Meta @scaleconference
Tuning Your Algorithm with MRS Memory System and Think-Then-Recommend (TTR) | Qi Guo from Meta  @scaleconference
Uploaded July 2026 | Updated November 2026, 2 days ago
Users have long wanted to understand and control the algorithms that shape their recommendations, but enabling meaningful user agency over recommendations has remained challenginging — until now. We present Tune-Your-Algorithm (TYA), an AI-powered agentic recommendation system on Instagram that gives users transparent visibility into the recommendation algorithm and the ability to tune it using natural languages.

TYA is built on two key innovations: (1) the MRS Memory System (Biography), an LLM-based framework that summarizes user engagement histories into rich, structured, and interpretable user interest and intent representations at scale; and (2) Think-Then-Recommend (TTR), a reasoning-augmented approach that decomposes user interests and complex user intents into personalized sub-goals for personalized and contextualized recommendations.

Early results show strong product-market fit with positive user feedback on transparency and user agency enabled by TYA. We discuss the end-to-end architecture, production learnings, technical challenges we are actively tackling, and the path towards our north star vision.

Learn more about the @Scale conferences here: atscaleconference.com
Tuning Your Algorithm with MRS Memory System and Think-Then-Recommend (TTR) | Qi Guo from MetaTo the moon and back!The Future of AI Infrastructure & Data Systems | @Scale 2026Building got easier. Judgment became scarce.What makes @Scale different?What happens when AI agents start working together?The 3-Step AI Evolution Every Creator Must KnowWhat does it take to build a network for AI at massive scale?@Scale: Product 2025A conference for engineers, by engineers!
@Scale |

Tuning Your Algorithm with MRS Memory System and Think-Then-Recommend (TTR) | Qi Guo from Meta

SHARE TO X SHARE TO REDDIT SHARE TO FACEBOOK WALLPAPER