Coming soon
Generative AI: LLMs, RAG and LangChain
Understand how LLMs work and build RAG apps that answer questions from your own documents.
- 2 months
- Intermediate (Python needed)
- 5 parts · 42 topics
Send us one message on WhatsApp, and we will tell you the launch date.
Planned syllabus
The full roadmap, basic to advanced
Every part builds on the one before it. You finish each part with a real project.
- Parts
- 5
- Duration
- 2 months
- Topics
- 42
- Projects
- 5
Part 1LLM foundations
- How LLMs are trained: pre-training and fine-tuning
- Tokens, context window, temperature and top-p
- Embeddings and vector similarity in plain words
- Choosing a model: Claude, GPT, Gemini, Llama, Mistral
- Calling the Claude and OpenAI APIs from Python
- Streaming responses
- Chat history and conversation memory
- Costs, token counting and rate limits
- Keeping API keys safe with environment variables
Project A command line chat app with memory.
Part 2Prompt engineering for developers
- System prompts and roles
- Zero-shot, few-shot and chain-of-thought prompting
- Structured output: JSON mode and schemas with Pydantic
- Tool use and function calling
- Prompt templates and variables
- Vision: reading images and PDFs
- Prompt caching to cut cost
- Prompt injection and how to defend against it
Project An invoice reader that turns a PDF into clean JSON.
Part 3RAG: answers from your data
- What RAG is and when you need it
- Loading PDFs, web pages and Word files
- Chunking strategies and chunk size
- Embedding models
- Vector databases: Chroma, pgvector, Pinecone
- Similarity search, filters and metadata
- Hybrid search: keywords plus vectors
- Re-ranking results
- Citing sources in every answer
Project Chat with your own PDFs and websites.
Part 4LangChain and evaluation
- LangChain basics: models, prompts, chains
- Retrievers and document loaders
- LlamaIndex: a quick comparison
- Building an evaluation test set
- RAG metrics: faithfulness and relevance
- LLM-as-a-judge
- Guardrails: stopping wrong or unsafe answers
- Tracing and monitoring with LangSmith
Project A support bot for a real product, with a test set of 50 questions.
Part 5Ship it
- A FastAPI backend for your AI app
- A simple chat front end
- Streaming answers to the browser
- Docker basics
- Deploying to the cloud
- Caching, rate limiting and cost control
- Logging user questions safely
- Writing a project README for your portfolio
Project Deploy your RAG app live and share the link.
This is the planned syllabus. Small changes are possible before launch.
Want this course? Tell us.
Send one message on WhatsApp, and we will tell you as soon as the launch date is fixed.