Inspiration
We wanted to build a system that could respond to manufacturing disruptions quickly while making decisions that are easy to understand and verify meaningfully
How we built it
We used Python, Qwen3 8B running locally through Ollama, and NetworkX for knowledge-graph traversal. Four specialized agents handle decisions, while deterministic code validates and executes actions through enterprise APIs. A React dashboard visualizes the results.
Built With
- javascript
- python
- tailscale
