A lightweight, offline-friendly, multi-agent tutor for HPC lesson planning, concept explanation, code scaffolding, execution, and review.
This repo implements the system described in the accompanying paper, centered on a Session agent that coordinates specialist agents (Explainer, Quizzer, Coder, Reviewer) and an Executor that compiles/runs code and returns observations for self-correction. It’s designed to work on laptops or institutional clusters without proprietary APIs.
Project status: early but usable. The code you see here is the minimal core. Additional folders (examples, jobs, scripts, tests, docs) are stubbed below for future population.
- Quick start
- How it works
- Repository structure
- Running on HPC systems
- Configuration
- Roadmap
- Contributing
- Security & privacy
- License & citation
- Acknowledgments
# clone
git clone https://github.com/epautsch/AgenticTutorHPC.git
cd AgenticTutorHPC
# python env (any of venv/conda/uv is fine)
python -m venv .venv && source .venv/bin/activate
# install deps
pip install --upgrade pip
pip install -r requirements.txt