Self-Service BI: Empowering Users with Luce
Learn how to implement self-service BI with Luce. Use semantic layers for non-technical users with Cube.js metrics and Looker integrations.
Read moreLuce is the consulting practice of Abdelkader Bekhti — AI solution architect, 10+ years building data and AI infrastructure for multi-billion-dollar companies, including a four-agent analytics platform for a global luxury group and the eval framework that proved it was safe to put in front of users.
The thesis behind the practice: companies are moving from human-operated to agent-first. The winners won't be the ones with the best demos — they'll be the ones who solve trust first: agents whose accuracy is proven, whose actions are bounded, and whose decisions can be audited.
Every engagement is senior-only: designed, shipped, and hardened by the person you talked to.
Most AI and data projects never reach production. I design and ship systems that survive contact with real users, real traffic, and real failure modes — and I measure success by what runs in production, not what demos well.

Abdelkader Bekhti
Founder & Principal Consultant
Field notes on data engineering, MLOps, and production AI — practical knowledge from real builds, not whitepapers.
Learn how to implement self-service BI with Luce. Use semantic layers for non-technical users with Cube.js metrics and Looker integrations.
Read moreLearn how to implement hybrid data architectures with Luce. Combine warehouses and lakes with BigQuery/S3 setups and DBT integrations.
Read more