My Projects
Explore my portfolio of Secure On-Premise AI solutions, MLOps implementations, and robust software architectures.
Secure On-Premise AI Infrastructure
Developed a robust, air-gapped AI environment integrating Large Language Models (LLMs) via Ollama and Open WebUI. Orchestrated complex local RAG (Retrieval-Augmented Generation) pipelines using n8n and PostgreSQL VectorDB for secure document analysis and automated summarization without cloud dependency.
Automated ML Evaluation & Scoring
Architected a production-grade machine learning solution utilizing CamemBERT for secondary-language (French) automated text evaluation. Built the full MLOps lifecycle including FastAPI endpoints, containerized deployment, and sophisticated retraining mechanisms triggered by performance drift.
Enterprise Management Platform
Led the development of a national administrative management system using a microservices architecture. Implemented secure RBAC (Role-Based Access Control) systems, high-performance Node.js/Express backends, and distributed data synchronization for large-scale governmental operations.
Operations Optimization (MIP)
Designed and implemented automated personnel assignment models using Mixed-Integer Programming (MIP). Leveraged PySCIPOpt and Pandas to solve complex combinatorial optimization problems, significantly reducing manual scheduling overhead through mathematical modeling.
MLOps Drift Detection Pipeline
Built a specialized MLOps monitoring system that automatically detects data and model drift in real-time. Created threshold-triggered retraining workflows in PostgreSQL to ensure AI model accuracy and reliability over time in changing environments.