GenAI Courses
Build real GenAI applications with AgentFlow. These courses teach you how to ship production-grade systems, not just explain LLM concepts in isolation.
Two tracks for different experience levels
GenAI Beginner
For software engineers new to LLM applications.
Go from "I know what an LLM is" to "I can build a production-shaped GenAI app with AgentFlow."
What you will build
A small engineer-facing assistant that:
- Answers using a curated knowledge source
- Uses one or two tools safely
- Accepts file or multimodal input
- Returns structured output
- Supports thread continuity or memory
- Streams responses to a client
- Includes a lightweight evaluation and release checklist
What you will learn
| Lesson | Topic |
|---|---|
| 1 | Use cases, models, and the LLM app lifecycle |
| 2 | Prompting, context engineering, and structured outputs |
| 3 | Tools, files, and MCP basics |
| 4 | Retrieval, grounding, and citations |
| 5 | State, memory, threads, and streaming |
| 6 | Multimodal and client/server integration |
| 7 | Evals, safety, cost, and release |
Prerequisites
- Python basics (functions, classes, async)
- Can read and write API request/response formats
- No prior LLM or agent experience needed
Time commitment
- 7 lessons × 30-45 minutes each
- 1 capstone exercise
- 1 shared release checklist
GenAI Advanced
For engineers who understand the basics and need to make reliable architecture choices.
Design runtime boundaries, choose between single-agent and multi-agent patterns, and prepare AgentFlow systems for production.
What you will learn
| Lesson | Topic |
|---|---|
| 1 | Agentic product fit and system boundaries |
| 2 | Single-agent runtime and bounded autonomy |
| 3 | Context engineering, long context, and caching |
| 4 | Knowledge systems and advanced RAG |
| 5 | Router, manager, and specialist patterns |
| 6 | Handoffs, human review, and control surfaces |
| 7 | Memory, checkpoints, artifacts, and durable execution |
| 8 | Observability, testing, security, and deployment |
Prerequisites
- Completed the GenAI Beginner course (or equivalent experience)
- Comfortable with AgentFlow core concepts
- Building or maintaining GenAI applications in production
Time commitment
- 8 lessons × 45-75 minutes each
- 1 architecture review exercise
- 1 production readiness worksheet
Shared foundations
Both courses depend on shared foundational concepts:
| Topic | Description |
|---|---|
| LLM basics for engineers | Mental model of what an LLM is, does well, and where it fails |
| Transformer basics | Enough architecture intuition to understand attention and context windows |
| Tokenization and context windows | Token budgets, prompt size, chunking, and cost reasoning |
| Embeddings and similarity | Vectorization, cosine similarity, and nearest-neighbor retrieval |
| Chunking and retrieval primitives | From embeddings theory to real retrieval systems |
| Prompt and output patterns | Reusable quick-reference for both tracks |
How the courses reinforce AgentFlow's value proposition
These courses teach you to:
- Start simple — pick the right use case before adding complexity
- Add tools and structured outputs — reliable interfaces between model and code
- Introduce memory and checkpoints — durable conversation state
- Grow into multi-agent only when needed — not every problem needs orchestration
- Ship with testing, safety, and deployment discipline — production-minded from day one
Draft
This curriculum is actively being developed. Share feedback in the AgentFlow GitHub repository.