Chabib El Malki

Founded xklic to generate and publish websites for small businesses using an agentic production line. AI agents generate websites in roughly 2–4 hours with automated checks and human approval, and the operation reached 50+ business clients within six months. Also built the operational workflow and delivery approach to support ongoing client onboarding and visibility.

Technical lead for an AI team of 4. Designed the architecture and built a self-hosted generative AI agent platform used daily by ~50 product owners and business analysts. Delivered an epic and user-story assistant that turns complex repository epics into user stories much faster, and automated code review triggered on every pull request (agent checks code against user story and recommended architectures). Implemented RAG over Confluence with internal governance, mandatory human approval for sensitive actions, controlled tool registry, prompt-injection protection, and full audit logs. Ensured production readiness with evaluation test sets (MLflow) and cost/error monitoring (OpenTelemetry) and deployed under bank security standards on OpenShift. Stack includes Python, Temporal, LangGraph/LangChain, Mistral, pgvector, MLflow, OpenTelemetry, Confluence/Jira, FastAPI, Docker, and GitLab CI.

Technical lead (freelance) in Corporate & Investment Banking. Built and operated a production AI data quality pipeline (2025–2026) that checks hundreds of thousands of market data files per day before ingestion, including column mapping, formats, anomalies, and missing values. Used a self-hosted open-source model (gpt-oss) with RAG on business rules/definitions when uncertain, achieving ~95% of corrections automatically; remaining cases were proposed for human approval. Tracked quality and reliability using LangSmith (self-hosted), DeepEval, promptfoo, and OpenTelemetry. Also built and maintained critical real-time and batch Go services between business teams and a central market data platform, plus reference data services with strict consistency/availability requirements. Modernized legacy Go services, standardized build/dependencies, and automated release promotion with Jenkins/Ansible including Jira/ServiceNow updates.