A multi-agent AI assistant that explores ESCO occupations and skills to support human-centered workforce restructuring.
This project uses LangChain, ChromaDB, and the ESCO taxonomy to simulate a collaborative AI system that assists with team planning after a company merger or restructuring. The system includes specialized agents (HR Strategist, ESCO Expert, Planner) and tools for semantic search, skill analysis, and career path planning.
- 🔎 Semantic search over ESCO occupations and skills
- 🧠 Skill gap analysis based on target roles
- 🗂️ Multi-agent simulation using LangChain
- 🧰 Custom tools for reasoning over structured data
- 💬 Dialogue-driven interface for realistic team planning scenarios
- Clone the repo:
git clone https://github.com/yourusername/teammerge-esco-agent.git
cd teammerge-esco-agent- Create a virtual environment and install dependencies:
pipenv install- Download the ESCO v1.2 datasets and place them in
./data/esco.
Open teammerge_agent.ipynb in Jupyter and run through the cells to:
- Load and index ESCO data with ChromaDB
- Define the agent roles and tools
- Simulate realistic restructuring conversations
Skill Search: Find ESCO skills semanticallyOccupation Search: Retrieve similar job rolesSkill Gap Analyzer: Compare current skills with target occupations
- Built with
LangChain - Uses
ChromaDBfor fast vector search - Embeddings via
sentence-transformers - Local ESCO data, no API calls required
This project is licensed under the MIT License — a permissive license that allows reuse, modification, and redistribution, even for commercial purposes, as long as the original license and attribution are included.
See LICENSE for details.
Built by Riccardo Di Sipio, PhD - Dayforce HCM. Inspired by practical questions about AI, work, and collaboration.