AI · Course 06 · 80 hours · 7 modules
TrendingGenerative AI & LLM Engineering
Build AI features that hold up in production.
A practical course on language models, retrieval, and agent workflows, focused on structured outputs, observability, and cost—not demos.
- Who it's for
- Developers and engineering teams adding LLMs and agents to real products
- Duration
- 80 hours
- Modules
- 7 modules, ending in a capstone project
- Formats
- Corporate training · College & university sessions · Own institute & online courses · Mentoring & internships
Syllabus
What the course covers
LLM fundamentals
- Tokens, context windows, and model selection
- Calling model APIs, including Azure AI Foundry
- Prompt design patterns
Structured outputs and tool calling
- Schema-constrained responses
- Retries, validation, and error handling
- Connecting APIs as tools
Retrieval-augmented generation
- Embeddings and chunking
- Vector search with FAISS and Qdrant
- Hybrid ranking with BM25
Agents with LangChain and LangGraph
- Stateful multi-step workflows
- Memory and checkpointing
- Multi-step execution and retries
Serving AI with FastAPI
- Designing AI-backed APIs
- API performance and response latency
- Persisting conversations and agent state
Evaluation and observability
- Tracing and evaluation with LangSmith
- Prompt and cost monitoring with Langfuse
- Token usage and latency tuning
Capstone project
- Design and ship a tool-connected agent
- Evaluate it against a test set before release
Interested in this course?
Share your audience, goals, and preferred format, and I'll tailor the structure.