OpsPilot — Case study
Personal upskilling project around incident management and FinOps, built with Java, Spring Boot and Next.js, with Kubernetes/Terraform examples kept at exploration level.
- Java
- Spring Boot
- Kubernetes
- Terraform
- Prometheus
- +1
I design and build web applications end to end: React and Next.js interfaces, Node.js APIs, PostgreSQL databases and Python automation. About three years of apprenticeship experience; open to a permanent role (CDI) or freelance missions, in Paris or remotely.

A selection of public projects, each one linked to its repository.
Personal upskilling project around incident management and FinOps, built with Java, Spring Boot and Next.js, with Kubernetes/Terraform examples kept at exploration level.
Personal upskilling project for real-time event access and ticketing with .NET, C#, Angular, SignalR and SQL Server.
A data science pipeline to predict Airbnb prices using Lyon and Paris data, with preprocessing, analysis, outlier handling, train/test split and linear regression models.
AI-assisted digital asset manager for food photography. Built with Flask and Python, it analyzes images via HuggingFace, generates tags and supports search/filtering workflows.
What I use professionally, what I build through personal projects, and what I only explore. The distinction is deliberate.
Technologies used at work and in repeated delivery.
Recent projects built to develop credible skills in new stacks.
Concepts and tools explored through examples, with no production claim.
Practical guides on Next.js, TypeScript, PostgreSQL and AI agents, written from my own projects.
How to structure a scalable SaaS application with Next.js 16, TypeScript and PostgreSQL: modular monolith, multi-tenancy with row-level security, authentication, billing, background jobs and production readiness.
17 min read
The real failure mode with Claude Code, Codex, Cursor, and Copilot is not prompt wording. It is context rot, tool access, weak permissions, and missing verification.
14 min read
A practical release playbook for moving an AI agent from proof of concept to controlled production workflow with permissions, evaluations, observability, go/no-go checks, and rollback.
7 min read
I'm a full-stack developer based in Île-de-France. Over about three years of apprenticeship I have worked mainly with JavaScript/TypeScript (React, Next.js, Node.js) and Python for automation and data workflows.
Alongside work I build personal projects to learn new stacks, most recently Java/Spring Boot and .NET/Angular, and I write about AI agents, MCP and developer tooling on the blog.