Benjamin Benoit

Technical Lead & GIS and AI Data Architect

Complete geospatial ETL pipeline

Public/private GIS data ingestion and data processing

PostGIS / PostgreSQL database management

Full Stack development of custom GIS web applications (Next.js + MapLibre + FastAPI) for local authorities, environmental consulting firms, architects, and land surveyors.

Automation of geomatic PDF reports (ReportLab + matplotlib), land identity cards, environmental studies, ecological compensation.

HD LiDAR and DTM analysis: rasterio/pysheds pipeline for 3D terrain models.

LLM integration for PLU regulatory extraction, interpretation of urban planning regulations, and local regulatory monitoring

Implemented a text-product semantic recommendation pipeline associating blog posts and corresponding products. Optimized search performance using clustering techniques based on embeddings and nearest neighbor search. Established data quality standards and restructured product descriptions. Designed an LLM-based evaluation framework for recommendation validation. Demonstrated significant improvement over keyword and TF-IDF methods through A/B testing.

AI Generative System and Voice Assistant Social Matching System

- Architect and developer of a complete Social Matching system based on a WhatsApp voice interface, using a multi-agent AI architecture to process natural conversations and generate personalized matches

- Deployment of infrastructure on Azure (Azure Functions, Azure OpenAI) with CI/CD integration via GitHub Actions for continuous deployment

- Implementation of a robust backend infrastructure combining specialized AI agents (interpretation, profile creation, matching). Design of a relational database (PostgreSQL) and a vector database optimized for efficient profile storage and semantic search.

- Setup of an integration automation system with Supabase and Azure Functions, using webhooks for profile vector updates