I build CI systems, autonomous developer loops, distributed Git infrastructure, and adaptive tools that turn complex engineering workflows into repeatable systems.
My work sits around one question: how much engineering coordination can be made explicit, testable, and runnable?
| Project | What it is |
|---|---|
| Titan | A modern CI engine in the spirit of Buildkite: pipeline execution, agent-driven work, and infrastructure-aware delivery. |
| vTrunk | A virtual-repo CI brain for teams that want monorepo-style CI without forcing all source code into one physical repository. |
| ForgeLoop | An autonomous engineering loop that turns well-specified GitHub issues into tested, reviewed, merged pull requests. |
| Pulsar | A distributed object plane for Git, designed around log-structured, deterministically named parts. |
| AdaptiQ | An AI-powered adaptive learning platform with generated practice, mastery tracking, and spaced repetition. |
| Retrobot | A modular event-driven automation system built around protocol decoding, state management, and workflow orchestration. |
| Confirmi | AI WhatsApp COD order confirmation for Moroccan e-commerce. |
- Modern CI engines and virtual monorepo workflows
- Autonomous development loops with typed review and merge gates
- Distributed Git storage, replication, and reachability
- Agentic learning systems and mastery tracking
- Event-driven automation with explicit state models



