Free course
Generative AI Full Course (Free)
A free generative AI full course you can follow on your own: what generative AI is, then a module-by-module syllabus from LLM basics to RAG, agents and deployment — each module with free material, a project and a checkpoint. Built for students in Pakistan and India with basic Python.
By Abdul Rahman Azam, founder of AI Season · Updated
Key facts
| Cost | Free — open-source code, free guides and free API tiers |
|---|---|
| You need | Basic Python and a laptop; no GPU and no machine-learning background |
| Modules | 8, from LLM APIs to deployed agents |
| Time | About 8–12 weeks part-time |
| Ends with | A portfolio of working generative AI projects on GitHub |
What is generative AI?
Generative AI is AI that creates new content — text, code, images, audio or video — instead of only classifying or predicting. Large language models (LLMs) such as GPT, Claude, Gemini and Llama generate text and code; diffusion models generate images. You use generative AI every time you ask a chatbot to write, summarise or explain something.
Agentic AI is the next step: generative models that plan and take actions with tools to reach a goal. This course covers both, because most real generative AI jobs now involve building agents. For the difference in depth, read what agentic AI is.
The free generative AI syllabus
| Module | What you learn | Free material |
|---|---|---|
| 1. How LLMs work | Tokens, context windows, temperature, why models hallucinate | AI agents glossary |
| 2. Calling an LLM from code | APIs, streaming, chat history, errors and cost | Session 1 code |
| 3. Prompt engineering | System prompts, few-shot examples, structured JSON output | Prompt engineering guide |
| 4. RAG over your own documents | Chunking, embeddings, vector databases, citations | RAG guide · Session 3 code |
| 5. Tool calling and AI agents | The agent loop, tools, guardrails | Build an AI agent · Session 4 code |
| 6. Frameworks | LangChain, LangGraph and no-code tools | LangChain tutorial · n8n AI agents |
| 7. Multimodal and MCP | Images and audio as inputs; connecting tools through MCP | Session 5 code · MCP guide |
| 8. Production | Evaluation, self-correction, deployment and monitoring | Session 8 code |
Want to see what is inside a language model? The optional ARGPT module in the free course material trains a small language model from scratch.
A project for every module
- A command-line chatbot with streaming and a switchable system prompt.
- A CV or invoice extractor that returns validated JSON.
- A question-answering bot over your university handbook that cites its sources.
- A tool-using agent — currency, weather or your own API.
- The same agent rebuilt in LangChain, plus an n8n version for comparison.
- An agent that reads an image or a PDF and acts on it.
- An evaluation suite of 30 test questions for your RAG bot.
- One project deployed to a public URL that a friend can use.
Put each project on GitHub with a short README explaining what it does and how you tested it. Eight small, working projects impress employers and clients far more than a certificate.
How to study it for free
- Models — use a provider's free tier or run a small local model; you do not need a paid plan to learn.
- Schedule — one module every 1–2 weeks, with the project finished before moving on.
- Order — follow the modules in order; RAG and agents are much easier once prompting and APIs are solid.
- Stuck? — read the error, check the model and key, and compare with the reference code before searching for a new tutorial.
For the complete learning path beyond generative AI, see how to learn AI and the AI agents roadmap.
Free course vs a live cohort
Self-study works if you are disciplined. What it cannot give you is live help when your code breaks, deadlines, and feedback on your projects. The AI Season bootcamp teaches this syllabus live — 12 sessions over 6 weeks, explained in Urdu, coded in English — for PKR 3,000 early-bird, with a full refund for the top 3 scorers. Cohort 02 starts 1st January 2027.
Frequently asked questions
Can I learn generative AI for free?
Yes. With basic Python you can follow a free syllabus using open-source course code, free guides and free model tiers, and build a portfolio of projects without paying for a course.
What is the difference between generative AI and agentic AI?
Generative AI creates content such as text, code and images. Agentic AI uses generative models to plan and take actions with tools towards a goal. Agents are built on top of generative models, so you learn generative AI first.
How long does it take to learn generative AI?
About 8–12 weeks of steady part-time study to finish this syllabus with basic Python, or 6 weeks in a structured live bootcamp.
Do I need maths or machine learning for generative AI?
No, not to build applications. Building with LLMs is mostly programming, prompting, retrieval, tool design and evaluation. Maths matters if you want to research or train models.
Is there a generative AI course in Urdu or Hindi?
AI Season teaches generative AI and AI agents live in Urdu and English, easy to follow for Hindi speakers, with all code in English.