Youssef El Beqqal
I help companies ship production AI systems: RAG pipelines, multi-agent architectures, and LLM workflows that actually work at scale.
I lead engineering teams building enterprise AI systems and high-traffic platforms serving millions of users. Based in Rabat, Morocco, working remote.
Work with me
I'm available for fractional AI engineering engagements and project-based work with US and European companies. If you're building a RAG system, deploying AI agents, or integrating LLMs into your product and need a senior engineer who has done this in production, let's talk.
About
I'm a Tech Lead with 7+ years building enterprise AI systems and high-traffic platforms at global scale. I currently lead technical delivery for a globally recognised sports platform serving millions of fans, and previously built production AI agent systems, including RAG pipelines, LLM-orchestrated chatbots, and automation workflows, for enterprise clients. I care about clean architecture, systems that hold up in production, and turning messy business requirements into things that actually work.
Experience
Technical Lead, Arsenal FC Platform · NTT DATA
Mar 2025 - Present
Leading full-stack delivery of a globally recognised sports platform serving millions of fans, from architecture through CI/CD.
AI Engineer, Automation CoE · NTT DATA
Jul 2024 - Feb 2025
Built production RAG pipelines and LLM-orchestrated chatbots using OpenAI and Claude, backed by Flask/Python APIs.
Sr. Full Stack Engineer, WIPO Sequence Tool · NTT DATA
Jun 2023 - Jun 2024
Built a cross-platform desktop and web application with Electron.js, Angular, and a Spring Boot backend.
Sr. Full Stack Engineer, HSE Home Shopping Europe · iMedia24 GmbH
Nov 2022 - May 2023
Built a solvency microservice in Kotlin/Vert.x on AWS CDK, and led a frontend migration to React.
Full Stack Developer · Skylark Services
Sep 2020 - Oct 2022
Delivered full-stack applications with React, Next.js, React Native, and Spring Boot.
Case Study
Multi-agent RAG system for a global sports platform
Arsenal FC's internal teams were spending too much time manually retrieving information across three separate business domains: customer support cases, historical match data for content editors, and internal policy documentation.
I designed and built a multi-agent RAG architecture where each domain has its own dedicated agent, retrieval pipeline, and vector store namespace. A central router directs queries to the correct agent using LangGraph. I moved the system from an initial monolithic agent design to this specialized architecture, which significantly reduced hallucination rates and improved response accuracy across all three domains. The system is deployed on AWS and monitored in production using LangSmith.
Skills
Frontend & TypeScript
Cloud & DevOps
Backend & AI