Skip to content

Projects & Engineering Highlights

This selection combines public open-source projects with professional platform-engineering work. Public repositories demonstrate how I approach architecture, automation, security and operational quality; client work is intentionally presented at a confidentiality-safe level.

Open-Source Projects

KubeLens

A lightweight Kubernetes topology dashboard and internal developer platform that helps teams understand workloads, services and networking relationships directly from the cluster control plane.
Technical highlights

  • Automatic local and in-cluster Kubernetes context discovery.
  • Least-privilege access through dedicated ServiceAccounts and RBAC.
  • DevSecOps quality gates covering linting, static analysis, IaC checks and container-image scanning.
Kubernetes Django Docker GitHub Actions Trivy

Explore KubeLens on GitHub

Multi-Tenant Airflow

An automated platform for provisioning multiple isolated Apache Airflow environments while sharing core PostgreSQL, Redis and routing infrastructure.
Technical highlights

  • Declarative tenant provisioning from YAML configuration.
  • Tenant isolation across databases, Celery channels and HTTP routes.
  • Automated validation of source code, configuration and generated Docker Compose infrastructure.
Apache Airflow Docker Compose PostgreSQL Redis Traefik

Explore Multi-Tenant Airflow on GitHub

Opsie

A cloud-native ChatOps engine that receives operational requests from Discord and processes them through an asynchronous FastAPI service with multi-model LLM routing.
Technical highlights

  • Decoupled, containerised Discord bot and backend API services.
  • Dynamic model-provider routing through LiteLLM.
  • Structured telemetry and automated CI quality and security checks.
FastAPI Discord.py LiteLLM Docker OpenAI Gemini

Explore Opsie on GitHub

Anatomopathology & Cancer Diagnostics

A Deep Learning project developed in collaboration with the University Cancer Institute of Toulouse (IUCT-Oncopole), focused on unsupervised analysis of microscopy images for cancer diagnosis.
Technical highlights

  • Designed feature-transfer and domain-adaptation approaches for microscopy image analysis.
  • Built unsupervised Deep Learning workflows for clustering and classification of medical images.
  • Developed reproducible Machine Learning pipelines with Python, TensorFlow, Keras, and Scikit-learn.
  • Containerised the development environment with Docker for consistent and portable experimentation.
Python TensorFlow Keras Scikit-learn Docker Unsupervised Learning Medical Imaging

Explore A.C.D on GitHub