I Built a Local RAG AI on My Laptop (No API Key Needed)| AI Systems | Aarti Dashore

Описание: How I Built a Local RAG AI on My Laptop (No API Key Needed)

This is an extended version of the semantic document retrieval system that I build which implements dotenv, CI/CD

GitHub Repo Link: https://github.com/AartiDashore/lab5

In this video I demo the RAG (Retrieval-Augmented Generation) system I built for Lab 6 of ARIN 5360 at Seattle University. It runs completely locally using Ollama — no OpenAI API key, no data leaving your machine.

What I am demonstrating:
`/health` endpoint confirming the LLM is live and ready
Full RAG query through my custom Playground UI — retrieving document chunks, building context, generating a grounded answer with source citations
All LLM and RAG tests passing in CI with zero real API calls (mocked with unittest.mock)
Honest reflection on what I learned building this

Tech stack:
🐍 Python + FastAPI
🗄️ ChromaDB (vector store)
🤖 Ollama (local LLM — Qwen 2.5)
🧪 pytest + unittest.mock
⚙️ GitHub Actions CI/CD

The core idea behind RAG:
Instead of relying on what the LLM was trained on, you retrieve relevant documents first, then ask the LLM to answer using only that context. Less hallucination. More accuracy. Full control.

This is the pattern behind most production AI apps — customer support bots, document Q&A tools, code assistants. Now I've built one from scratch.

---
🎓 Course: ARIN 5360 — AI Systems: Document Retrieval, Seattle University

---

#RAG #AI #Python #Ollama #MachineLearning #FastAPI #LocalAI #LLM #ComputerScience
#AI #SemanticSearch #FastAPI #PythonProject #MachineLearning
#VectorDatabase #cicd #ChromaDB #SentenceTransformers #AIEngineering
#SoftwareEngineering #BackendDevelopment #APIDevelopment
#StudentProject #CSStudents #AIProjects #Pytest #Python #AIProjects #GoogleAlgorithm #SearchEngine #RAG #InformationRetrieval #CrossEncoder #BiEncoder #NLP #DeepLearning #MLOps #bm25 #searchengineoptimization #searchengineoptimizationservices #searchengineranking