← Back to Work

Consumer Intelligence

AI Taste Mapping Platform

Built the AI infrastructure for understanding and predicting consumer preferences across 3.7B+ lifestyle entities.

ML Infrastructure Data Engineering Production Systems

Key Outcomes

  • Scaled to 3.7B+ entity knowledge graph
  • Real-time preference predictions
  • Privacy-preserving taste modeling

The Challenge

How do you map human taste at scale? Qloo needed an AI system that could understand and predict consumer preferences across music, food, fashion, travel, and more—without compromising user privacy.

The Approach

I led the development of cultural intelligence infrastructure that connects billions of lifestyle entities through learned taste relationships. The system uses collaborative filtering, knowledge graph embeddings, and contextual signals to understand what people like and why.

Technical Highlights

  • Designed scalable ML pipelines processing billions of preference signals
  • Built knowledge graph with 3.7B+ entities across lifestyle domains
  • Implemented privacy-preserving recommendation algorithms
  • Architected real-time inference systems for sub-100ms predictions

The Outcome

The platform now powers taste-based recommendations for major brands, helping them understand their customers without invasive data collection.

Interested in Similar Work?

Let's discuss how I can help with your AI project.

Get in Touch