🚀 Generative AI & RAG with LangChain 1.x

This repository contains end-to-end implementations of Generative AI and Retrieval-Augmented Generation (RAG) systems built using LangChain 1.x, following the modern runnable-based pipeline approach instead of deprecated chain abstractions.

The project demonstrates a complete real-world GenAI workflow — from data ingestion to embeddings, vector stores, retrievers, and context-aware LLM responses, using both OpenAI and open-source models (Ollama, HuggingFace).

---

🧠 What You Will Learn

• Modern LangChain 1.x runnable pipelines
• Retrieval-Augmented Generation (RAG) from scratch
• Vector similarity search using FAISS & ChromaDB
• Context-aware LLM generation
• Working with both cloud (OpenAI) and local (Ollama) models
• Debuggable, production-ready GenAI architecture

---

🛠️ Tech Stack

• LangChain 1.x
• OpenAI (GPT-4 / GPT-4o-mini)
• Ollama (Local LLMs & Embeddings)
• FAISS & ChromaDB (Vector Stores)
• HuggingFace Embeddings
• Python
• Streamlit (Chat UI)
• LangSmith (Tracing & Observability)

---

✨ Key Highlights

• ✅ Uses pure LangChain 1.x runnable pipelines
• ❌ No deprecated langchain.chains APIs
• ✅ Explicit retrieval and context injection
• ✅ Supports cloud + local LLMs
• ✅ Clean, debuggable, production-style code
• ✅ Streamlit-based chatbot interface
• ✅ LangSmith tracing for observability

📌 Author Nikhil Shukla Generative AI | LangChain | RAG | LLM Engineering