Internal tools and deployment
I was a Data & AI Engineer on a permanent contract at Call To Action (Bonum Group), in Nice, from mid-August to 16 October 2026, working on internal tools, data and the deployment of those tools.
Guarded writes on SQL Server
- A Python tool, tested in a sandbox (a trial environment), that changes the database in short batches, with a plan reviewed before it runs and a possible rollback. It has 15 safeguards tested automatically, with GitHub Actions continuous integration on SQL Server 2019 and 2022.
Applications in containers
- Callytics: the FastAPI backend and Next.js frontend moved to Docker, alongside the running instance; GitHub Actions CI/CD (linting, build, secrets); deployment and rollback scripts on the server.
- StatsLive, taken over with a colleague: refactoring, GitHub Actions CI/CD back to green (linting, tests, build, secrets; 6 jobs), central session check on the API routes, Docker image and automated deployment to the server.
Tools for the IT team
- Self-hosted Redmine 6.1: ticketing tools compared, then a container deployment kit (Docker Compose, PostgreSQL, custom theme), a restored backup, a 10-point acceptance run. Deployed and accepted, not yet open to the whole team. A team project; the kit is mine.
- A monitoring prototype (Next.js, React, TypeScript, in a Docker container): a working mock-up with 48 indicators coded.
- A local AI assistant (Open WebUI, Ollama, RAG over the knowledge base), installed on the local network for the IT team. I have no record of its use.
- A knowledge base of the information system for coding agents: about 120 Markdown files, with read-only Python scanners.