"The Cognitive Revolution" | AI Builders, Researchers, and Live Player Analysis Milliseconds to Match: Criteo's AdTech AI & the Future of Commerce w/ Diarmuid Gill & Liva Ralaivola
39 snips
May 9, 2026 
Liva Ralaivola, VP of Research at Criteo, builds embeddings and foundation models for real‑time recommender systems. Diarmuid Gill, CTO at Criteo, leads engineering, privacy‑first ad‑tech and partnerships. They discuss millisecond recommendation systems, embeddings and foundation models, privacy‑aware personalization, real‑time bidding architecture, and generative creative for commerce.
AI Snips
Chapters
Transcript
Episode notes
Provide Transparent Opt Outs To Preserve Trust
- Be transparent and offer opt-outs to maintain user trust in personalized advertising.
- Criteo pioneered the AdChoices icon so users can see why they saw an ad and opt out, preserving the value exchange that funds free content.
Hybrid LLMs Need Fresh Commerce Feeds
- LLMs bring rich reasoning and conversational discovery but lack fresh commerce data.
- Criteo ingests product feeds from 17,000 retailers multiple times daily, so hybrid LLM + live inventory yields accurate recommendations.
Many Foundation Models Power Reusable Embeddings
- Criteo builds multiple foundation models that output reusable embeddings for products, users, and timelines.
- Those embeddings serve as hot-starts across teams, let engineers warm-start models, and are versioned like software to avoid breaking downstream systems.