Real-Time Fraud Detection Pipelines
How to build real-time fraud detection pipelines using Kafka streaming, DBT for pattern detection, and Cube.js for metrics. Production architecture achieving 15% fraud reduction.
Read moreAbout - Building Systems That Work in Production
I've spent 10+ years building the data infrastructure that powers multi-billion dollar companies. Now I help growth companies build the same enterprise-grade systems—without the enterprise overhead.
Throughout my career, I've built platforms that process over 1 billion events per day, led teams of 10+ engineers, and designed cloud-native data architectures for some of the world's most demanding luxury brands and financial institutions. I've seen what works at scale—and what doesn't.
I've also founded production AI companies, including Nestorchat, my suite of 4 live AI assistants with real-time voice processing. I believe in building, not just advising. Every system I design is built to actually work in production, not fail at scale.
Today, I'm based in Dubai and work with select growth companies who need enterprise-scale systems but can't afford enterprise timelines. I limit myself to 3-4 projects per year to ensure exceptional quality and deep involvement in each engagement.
- Years building data platforms
- 10+
- Events/day processed
- 1B+
- Projects per year (limited availability)
- 3-4
How I work - Production-ready systems, not prototypes that fail at scale.
I've built systems at billion-event scale. I know what it takes to go from prototype to production—and I build that in from day one.
- Built for Scale. Every architecture I design is built to handle enterprise-scale loads. I've processed billions of events per day—I know how to build systems that don't break under pressure.
- Production from Day One. I don't build prototypes that fail in production. From MLOps pipelines to real-time streaming, everything is designed to actually work at scale, with monitoring and observability built in.
- Limited Availability. I limit myself to 3-4 projects per year. This isn't a side gig—when I take on a project, I'm deeply involved from architecture design through production deployment and beyond.
About
Abdelkader Bekhti
Founder | Production AI & Data Architect
Latest insights
Technical deep-dives on building production-ready AI and data systems at scale. Real-world lessons from billion-event platforms and production ML deployments.
Building a Data Mesh: Lessons from Retail
How to implement a decentralized data architecture, scaling to 10 domains in 8 weeks using domain-driven DBT models and Terraform automation. Real-world lessons from retail.
Read moreReady to build production-ready systems?
Based in Dubai
- Dubai
Dubai, UAE
Currently accepting limited engagements