Abstract
It’s easy to look at a mature AI platform and imagine a grand blueprint. Ours began with none. What started as a few data scientists hacking on R scripts has grown into an AI platform serving millions of predictions and decisions every day. This is the story of that transformation from an engineering perspective.
This talk traces Booking.com's AI architecture evolution over the last two decades—from deterministic formulas and cron jobs to centralized ML platforms for Ranking, Recommendations, Content Intelligence and GenAI. We'll focus on the backend and infrastructure choices that made that journey possible: data pipelines, serving layers, cloud migration, and platformization. No data science deep dives—just real engineering trade-offs, missteps, and inflection points behind the current AI Platform.
In this presentation, you'll learn from our two-decade journey, with particular focus on:
- Architecture patterns and trade-offs for serving millions of AI predictions at scale
- Real lessons learned from data pipeline design, feature engineering and serving layer implementation
- The organizational and technical inflection points that shaped our platform evolution
- Common pitfalls to avoid and practical patterns that worked building AI at scale
Speaker
Jabez Eliezer Manuel
Senior Principal Engineer @Booking.com - Building Next-Gen AI Platform, 20 Years in Software Engineering
Jabez Eliezer Manuel is a Principal Engineer @Booking.com with two decades of Software Engineering Experience.
Jabez specializes in mission-critical engineering. At Booking.com, he’s spent a decade tackling high-scale challenges—from delivering high-throughput search personalization in milliseconds to designing the company's unified payment ledger and most recently, a long-term memory system for AI agents.
Today, he drives the engineering behind the AI Application platform that helps product teams ship intelligent and compliant products at scale.