Self-paced · Hands-on · Colab-style

Build AI agents,
one lesson at a time.

A complete hands-on course for developers learning to build with LLMs. Read the explanation, copy the code, run it on your own machine, and reveal solutions whenever you get stuck.

9
Sessions
90+
Lessons
4
Projects
Self
Paced

How the portal works

3 simple steps per lesson
1
Read the lesson
Each exercise has a short explanation of what the code does and why it matters.
2
Copy & run locally
Click copy on any code block, paste into your editor, run from the terminal.
3
Check yourself
Knowledge-check quizzes, hint reveals, and full solution toggles when you need them.

The learning path

9 sessions, in order
0
3-4 hours

Fasttrack — Complete AI/LLM Crash Course

A visual, self-contained super tutorial: every core concept from your first LLM call to multi-agent orchestration and evals, in 10 lessons.

LangChainProvidersStructured outputMemoryTools
1
1 hour

Session 0 — Foundations

Understand LLMs, providers, costs, and how to choose a model before you write code.

LLM basicsML vs LLMProvidersCostModel choice
2
2.5–3 hours

Session 1 — OpenAI SDK Fundamentals

Text, vision, image generation, and text-to-speech from scratch.

Text generationVisionDALL·ETTSCLI apps
3
3 hours

Session 5 — Structured Outputs & Function Calling

Force the model into JSON, call tools, and orchestrate multi-tool flows.

JSON modeJSON SchemaPydanticTool callingRefusals
4
60–90 minutes

Session 6 — LangChain Multi-Provider

Same code, three providers: OpenAI, Anthropic, and Gemini.

LangChainOpenAIClaudeGeminiStreaming
5
1 hour

Session 7 — Choosing Agent Architecture

When to use workflows vs agents vs multi-agent — pick the right pattern before you code.

WorkflowsAgentsMulti-agentArchitectureDecision framework
6
3 hours

Session 8 — Building Agents with LangChain

Build your first agent, add memory, streaming, structured output, and multi-agent handoffs.

create_agentReActMemoryStreamingHITL
7
2–3 hours

Session 9 — LLM Evaluations

Build eval datasets, score outputs with LLM-as-Judge, and generate HTML reports.

EvalsLLM-as-JudgeDatasetsHTML reportsAdversarial testing
8
1.5–2 hours

Session 10 — Prompt Engineering

Improve eval scores with disciplined prompt iteration: specify, structure, constrain, align.

Prompt engineeringA/B testingSpecifyConstrainFew-shot
9
3 hours

Session 11 — RAG (Retrieval Augmented Generation)

Chunk documents, embed them, build a vector store, add BM25, and answer questions with Claude.

ChunkingEmbeddingsVector DBBM25Hybrid retrieval

What you'll have built

Real artifacts — not just toy snippets — that you can run, show off, and extend.

💬
A multi-modal CLI app
Summarise, analyse images, generate art, and speak — all from one Python script you build in Session 1.
🛠️
A tool-calling agent
Define Python functions, let the LLM call them, and build a loop that handles errors and parallel requests.
🏨
A real booking agent
The BookingAgent capstone — LangChain + SQLite, a full hotel reservation chatbot with persistent data.
🎤
A full-stack interview bot
InterviewerBOT — React frontend, Express backend, OpenAI streaming, end-to-end architecture.

Capstone Projects

Built for self-paced study · Ask questions in your session Slack · Code lives in the source folders alongside this portal.
Get started →