Materials for the DataHack Summit 2026 workshop by Alessandro Romano.
Build a production AI Knowledge Assistant — from a single LLM call to a deployed, multi-tool agent with RAG, memory, and MCP.
- 📊 Slides — view the deck
(source in
web/; pressFfor fullscreen while presenting) - 📁
src/— all the learning materials: hands-on notebooks (per module) and the Knowledge Assistant app you build and deploy. Seesrc/README.mdfor setup and how to run.
cd src
uv sync # install everything# add your API keys to src/.env
uv run jupyter lab # run the notebooks
uv run streamlit run app/solution/app.py # run the appMIT — use, adapt, and teach from this freely.
Bundled third-party code keeps its own license: web/vendor/ contains
reveal.js 5.1.0, also MIT, © Hakim El Hattab.