README.md

RAG 30 Days Sprint 🚀

This repository contains a 30-day sprint to master Retrieval-Augmented Generation (RAG) systems using Python, LangChain, and modern AI tools.

📅 Day Tracker

Day	Folder	Description	Status
1	day1	Hello world test file	✅
2	day2	TBD	⏳
...	...	...	...

📂 Folder Structure

rag-30-days/ │ ├── day1/ │ └── hello_ai.py │ ├── README.md

markdown Copy Edit

🧠 Goal

To build a production-ready RAG pipeline in 30 days and land a remote AI job by the end of the sprint.

🛠️ Tools

• Python
• LangChain
• ChromaDB / Weaviate / FAISS
• OpenAI API
• Streamlit (optional UI)
• Git & GitHub

📈 Progress

Check commits and folders daily to follow the sprint. Each folder corresponds to 1 day of learning and building.

📅 Day 1 – Getting Started with Python & Flask

✅ What I Learned

• Refreshed core Python basics (variables, functions, classes, etc.)
• Built my first Flask API with real-world JSON responses
• Practiced structured coding with Copilot assistance

🛠️ What I Built

• hello_ai.py: A minimal Python script to print a welcome message
• api.py: A Flask application with 3 endpoints:
  • /hello: greeting message
  • /calculate: accepts 2 numbers (POST) and returns their sum
  • /ai-ready: motivational message for AI learning

🔮 Tomorrow's Plan

• Begin LangChain setup and environment configuration
• Start working on RAG-based document processing
• Set up folder structure and day2 workflow

👣 One day down, 29 to go. Keep shipping.

Day 3: First RAG System ✅

What I Built

• PDF processing pipeline (loader + optimal chunker)
• Compared 3 chunking strategies (fixed, recursive, token)
• ChromaDB vector storage (persistent)
• SentenceTransformer embeddings (MiniLM)
• Gradio chat interface (upload PDF → ask)
• Deployment on Hugging Face Spaces

Key Learnings

• Fixed vs Recursive vs Token-based chunking trade-offs
• Embedding format must be list[list[float]] for Chroma
• New Chroma API uses PersistentClient
• Prompt design: extractive answers + fallback

Live Demo

🔗 HuggingFace Space Link

📬 Contact

Made by Hamid Omarov
Check out my portfolio: Notion Page