Idea · Wed, 28 Oct, 2026 · 07:49 am UTC
AI Engineering Assistant
Engineers often spend time searching documentation, understanding unfamiliar codebases, debugging recurring issues, and asking teammates questions that have already been answered.
Proposed solution
Build an internal AI assistant that understands our code repositories, technical documentation, architecture decisions, and resolved bugs. Engineers can ask questions in natural language and receive answers with relevant source links.
Suggested technology / approach
LLMs, RAG, embeddings, vector database, GitHub integration, Slack integration, and a Node.js API.
How to contribute
I'd like to build it myself.
Notes
Start with one repository and its documentation, measure answer accuracy and time saved, and expand based on results.
Submitted by
Amit Patil
