Recommender-system of the OpenSource Together platform.

---

What is it?

An AI‑powered data pipeline that discovers, understands, and curates open‑source projects to power OST’s recommendation system and provide high-quality projects to contribute on.

What it does :

• Discover: scan GitHub at scale with Golang scrapers
• Understand: detect language and semantics (fastText + transformers)
• Assess: score quality and relevance from activity and metadata signals
• Enrich: normalize topics, tech stacks, and fields into a coherent schema

Deliver: output a clean, queryable dataset (PostgreSQL via Prisma)

Quick Start

1. Copy .env.example into .env and fill it.
2. Copy config/cfg_example.py to config/cfg.py and adjust the config to your personal parameters.
3. Start

# Start the engine
docker compose up

Dagster UI : http://localhost:3000

Status

Work in progress.
Build in public here : @spideyX

---

Made with love by @spideyX & the OST team for the OSS community