← Retour

Tous les projets

Vue complète des projets en vedette et projets personnels.

Projets en vedette

DocPipe

$ langgraph main.py

~€0.05 / doc · in production at GYS

DocPipe

Production LLM pipeline turning industrial PPTX/PDF/RST docs into clean, diagram-ready Markdown for GitLab.

~€0.05 / doc · in production at GYS

LangGraphOpenRouter/Gemmapython-pptxMarkerLangSmith

Hermes Agent

$ langgraph main.py

Live / operational

Hermes Agent

Multi-agent homelab orchestrator I drive from Telegram — routes tasks to specialized sub-agents.

Live / operational

LangGraphGeminiTelegramMCP

GDC — Firmware Update System

$ stm32 main.py

9 min → 2 min 20 (−74 %), throughput ×5

GDC — Firmware Update System

Automated firmware-update system for a real industrial product line, over CAN bus on STM32.

9 min → 2 min 20 (−74 %), throughput ×5

STM32FatFsCAN/CAN-FDDMABootloader

Autres projets

Démo

Clothify

Discord bot generating product visuals from user selections via n8n workflow + Gemini image model.

Dynamically built prompt from body type, size, angle parameters

Discord.jsn8nGeminiPython
Démo

Multimodal SNCF Routing

Shortest-path over the French rail network: NeTEx/GTFS XML → Neo4j graph, Dijkstra & A*.

Full French rail graph imported and queryable

PythonNeo4jGraph Data Sciencepandas
Démo

Collaborative Delivery Optimization

VRPSPDTW — multi-hub parcel exchange between cargo-bike couriers with time windows.

Models pickup & delivery with time windows across multiple hubs

OR-ToolsPythonHeuristicsOptimization
En cours

n8n Voice Agent

Personal orchestration assistant: voice notes capture & classification, semantic router, proactive calendar, smart email reading.

Full automation of personal knowledge management

n8nWhisper (STT)EmbeddingsGoogle Calendar APIGmail API
Démo

High-Frequency Scraper

Async high-throughput scraping with anti-bot handling and proxy rotation.

Production-grade pipeline handling large-scale data extraction

PythonPlaywright/BotrightAsyncIORedis/MongoDBProxy rotationCelery
Démo

Multi-Agent LLM Pipeline

Multi-agent prompt/processing orchestration with user classification and structured output.

Composable agent graph with JSON-Schema validated outputs

AutoGenJSON-SchemaPython