Generative AI & LLM Engineering: Build Real AI Apps

What you'll learn

• Write reliable prompts and structure LLM outputs your application can actually depend on
• Build a Retrieval-Augmented Generation system that answers from your own documents
• Use embeddings and a vector database to give models real, up-to-date context
• Design and build a tool-using AI agent that completes multi-step tasks
• Decide when to fine-tune versus prompt, and run a fine-tune end to end
• Evaluate and guard LLM outputs so your feature is safe to put in front of users

About this course

Everyone can call an LLM API. Far fewer can build an AI feature that's reliable enough to ship — and that's exactly the gap this course closes. It's aimed at developers who want to move from impressive demos to production-grade generative-AI systems. You'll start with prompt engineering that treats the model as an unreliable component to be constrained, not a magic box: structured outputs, system prompts, and evaluation. Then you'll build Retrieval-Augmented Generation from scratch — chunking, embeddings, a vector database, and grounded answers over your own data — the single most useful pattern in applied LLMs today. From there you'll build a tool-using agent that plans and executes multi-step tasks, learn when fine-tuning genuinely helps (and run one), and finish with the part most tutorials skip: evaluating outputs, catching hallucinations, and adding guardrails. You should be comfortable with Python and APIs. By the end you'll have built a real RAG application and an agent, and you'll understand the engineering decisions behind dependable AI products.