TASKS.md

Workshop Tasks

Complete these tasks in order to gain a comprehensive understanding of LangChain, RAG, and LLMs.

1. Setup and Introduction

• Install dependencies and set up TypeScript configuration
• Create a simple "Hello, World!" script using LangChain with TypeScript

2. LangChain Basics

• Implement a basic LLM chain using TypeScript
• Create a custom PromptTemplate and use it in a chain

3. Working with Vector Stores

• Set up a local vector store (e.g., HNSWLib) with TypeScript
• Implement document loading and splitting

4. Building a Simple RAG System

• Create a TypeScript class for a basic RAG system
• Implement retrieval and generation steps

5. Advanced RAG Techniques

• Implement a conversation memory system
• Create a custom retriever using LangChain and TypeScript

6. LangSmith Integration

• Set up LangSmith for tracing and debugging
• Implement logging and monitoring for your RAG system

7. Using Hugging Face Models

• Integrate a Hugging Face model into your LangChain application
• Implement a custom embedding model using Hugging Face Transformers

8. Building a Complete Project

• Design and implement a RAG-based Q&A system
• Add error handling and input validation

9. Testing and Optimization

• Write unit tests for critical components
• Implement performance optimizations (e.g., batching, caching)

10. Deployment and Best Practices

• Containerize your application using Docker
• Implement proper environment variable handling and security best practices

Bonus Challenges

• Implement a hybrid search-and-generate system
• Create a multi-modal RAG system (text + images)
• Build a simple web interface for your RAG application