Building with LLMs
For software engineers who can call an LLM API and now need to ship a real feature around it. You will compose chains, tools and agents, choose a provider and model tier by cost, handle memory, structured output and external APIs safely, and deploy a document Q&A service with retries, caching and streaming.
Course Content
Shows when a problem needs a chain, tools or an agent, how to choose a provider and model tier, and how to build pipelines you can test and swap.
Covers the three jobs every real LLM feature has: remembering a conversation, calling outside systems safely, and getting typed data in and out of the model.
Takes an LLM service to production: secrets handled safely, retries and caching that survive an outage, and an async API you can run in a container.
Puts the course together in one build: a document Q&A service with retrieval, citations, an honest "I don't know", an API, a UI and a container.