Transformers at scale to boost recommender systems - Mirakl at dotAI Tech Track AI Day

Why watch

This talk presents Mirakl's real-world implementation of transformer-based recommender systems for their ad platform. The speakers explain how they used short-term user behavior, reconstructed sessions, and learned product embeddings to improve sponsored product relevance at marketplace scale. Ideal for ML engineers and data scientists working on production recommender systems who want to learn from a concrete case study.