This repository contains the development of an AI-powered knowledge assistant that helps users navigate technical documentation using Retrieval-Augmented Generation (RAG). The project is being built over a structured 10-week plan and explores technologies like LLMs, vector databases, LangChain, and Streamlit — with a focus on applications in technical enablement, onboarding, and knowledge management.
To build a hands-on RAG-based prototype that:
- Retrieves relevant knowledge from technical documentation
- Generates contextual, verifiable responses using LLMs
- Adds learning-focused features like guided flows, concept linking, and embedded knowledge hygiene tips
- Demonstrates how modern AI tools can support documentation-driven learning at scale
- Language: Python
- LLMs: Azure OpenAI / OpenAI GPT-4
- Vector DB: ChromaDB (local), Pinecone (cloud), or Azure Cognitive Search
- Frameworks: LangChain / LlamaIndex
- Interface: Streamlit or Gradio
- Other: Docker, GitHub Actions, RAGAS (evaluation)
This project is being developed across a 10-week learning sprint, with weekly milestones and blog posts published along the way.
| Week | Focus Area |
|---|---|
| 0 | Foundations, Intentions & Direction |
| 1 | Project Scope, Setup, Data & Market Research |
| 2 | Document Parsing & Embeddings |
| 3 | Vector Store Setup & Retrieval |
| 4 | LLM Integration & Basic RAG Q&A |
| 5 | Refining Retrieval + Prompting |
| 6 | Evaluation Framework |
| 7 | Learning-Focused Features |
| 8 | UI Build (Streamlit/Gradio) |
| 9 | Dockerization & Deployment |
| 10 | Final Polish & Strategy Write-Up |
This assistant is designed to help technical and non-technical employees — especially new hires — quickly locate relevant information across sprawling internal documentation repositories (e.g., Confluence, SharePoint, GitHub wikis), while offering embedded micro-learning and support.
Weekly blog posts will be published on JenShannon.tech and linked here once available:
- Week 0: Beyond the Playlist: Rethinking Onboarding with Embedded Intelligence
- Week 1: Coming soon…
This project is a learning experiment developed in public. I’m building as I learn — expect rough edges, questions-in-progress, and plenty of iteration. Contributions and feedback are welcome (especially if you're also exploring AI + knowledge systems)!
This project is licensed under the MIT License. Feel free to fork, build on, or contribute — attribution appreciated.
Hi, I’m Jen Shannon, a technical learning and enablement leader passionate about AI, knowledge management, and workforce development. I'm using this project to explore how we might bring learning and documentation closer together in modern organizations.
Let’s connect: