AI and LLMs for Absolute Beginners
The fastest, most practical, 2026-updated way to go from zero to building real AI apps.
TL;DR: 12 modules. 40 to 60 hours. No prior coding required. By the end you will understand how ChatGPT works, write production-grade prompts, use the OpenAI and Gemini APIs, build a PDF chatbot powered by RAG, and deploy it publicly.
Why this course exists
Most "AI courses" either drown you in math or sell you hype. This one is different. It is built like a real engineering apprenticeship: explain the idea in plain English first, then go technical, then show how it is used in production, then make you build it.
You will not just learn what a token is. You will count tokens, estimate cost, and ship an app that streams tokens to a real user. You will not just hear about RAG. You will build a PDF chatbot that uses it.
Who this course is for
- Complete beginners who have never written code
- Students wanting a career edge before placements
- Working professionals planning a switch into AI
- Non-technical founders who want to build AI products
- Developers from other stacks who never formally learned LLMs
If you can use a browser and install an app, you can finish this course.
What you will learn
By the end of this course, you will be able to:
- Explain how ChatGPT and modern LLMs work, in plain English and at a technical level.
- Write prompts that consistently produce reliable, structured outputs.
- Use the OpenAI, Gemini, and Anthropic APIs from Python.
- Build a working AI chat app with streaming responses.
- Understand and implement Retrieval Augmented Generation (RAG).
- Build a PDF chatbot that answers questions from your own documents.
- Deploy AI apps to the public internet on Streamlit Cloud, Vercel, or Hugging Face Spaces.
- Estimate and control API costs in production.
- Identify hallucinations, prompt injection, and AI safety risks.
- Plan your next 6 months as an AI engineer with a clear roadmap.
Career outcomes
This course is the on-ramp to several real, hireable roles in 2026:
| Role | Typical entry path | What this course covers |
|---|---|---|
| AI App Developer | Beginner to junior | Modules 1 to 10 fully |
| Prompt Engineer | Beginner | Module 4 deeply |
| Generative AI Engineer | Intermediate (next step) | Foundation for it |
| AI Product Manager | Non-engineering | Modules 1, 2, 3, 11, 12 |
| AI Educator / Creator | Any | Whole course as basis |
| RAG / AI Search Engineer | Intermediate (next step) | Modules 7, 8, 9 |
Resume-ready skills you will gain
- Python basics for AI
- OpenAI, Gemini, and Anthropic API integration
- Prompt engineering patterns (few-shot, chain-of-thought, system prompts, JSON mode)
- Vector embeddings with OpenAI and Hugging Face
- Vector databases (ChromaDB, FAISS)
- RAG pipelines from scratch and with LangChain
- Streamlit for AI UIs
- FastAPI for AI backends (light intro)
- Deployment on Streamlit Cloud, Vercel, Hugging Face Spaces
- Token accounting and API cost control
- Responsible AI and hallucination mitigation
Course structure at a glance
Plus: Hands-on projects and capstone, Tech stack and tools, and Assessments, certificates, and XP.
Prerequisites
- A laptop or desktop with at least 8 GB RAM
- Stable internet
- A free Google account (for Colab) and a free GitHub account
- Curiosity. That is genuinely the only one that matters.
Optional: 30 minutes of basic Python familiarity. If you have none, Module 5 includes a "Python for AI in 30 minutes" primer.
How to learn from this course
- Read the lesson fully first. Do not skim. Lessons are tight, but every paragraph carries weight.
- Do the Hands-on Lab before the Quiz. Building cements the idea.
- Submit your project link in the community thread. Public commitment doubles completion rates.
- Skip the Challenge Tasks on the first pass if you are stuck. Come back after Module 6.
- Use AI to learn AI. When stuck, ask ChatGPT or Claude to explain the same idea three different ways.
Visual learning suggestions
This course is designed text-first but visually immersive. Each module suggests where diagrams should appear (architecture, flowcharts, token visualizations). If you learn better with visuals, sketch them yourself in a notebook as you read. The act of drawing locks the concept in.
Gamification
- XP: Every lesson, quiz, and project awards XP.
- Streaks: Daily learning streaks unlock badges.
- Tiers: Bronze (Module 1 to 4), Silver (5 to 8), Gold (9 to 12), Platinum (capstone shipped).
- Leaderboard: Course leaderboard updates weekly. See Assessments and certificates.
Certificate
Finish all 12 modules, pass the assessments, and ship the capstone. You will receive a verifiable GeekHub certificate with a public proof URL and the GitHub link of your capstone. See criteria in Assessments.
Common beginner mistakes this course prevents
- Reading 50 hours of theory before writing a single prompt
- Memorizing models that will be obsolete in six months
- Treating prompts as magic spells instead of instructions to a junior employee
- Building chatbots on top of LLMs without thinking about hallucinations
- Spending money on API calls without watching the token meter
- Picking outdated libraries because a 2023 YouTube tutorial said so
What this course is not
- Not a math course. We touch the minimum required intuition.
- Not a fine-tuning course. That is Module 1 of the intermediate course.
- Not a survey of every AI tool ever made. We pick the best and go deep.
- Not a quick "make money with AI" hack. It is real engineering.
Frequently asked questions
Is this course free? Yes. Every lesson, lab, and project is free. The certificate is free too.
Do I need to know Python? No. From Module 5 we introduce just enough Python for AI. Beginners finish it with comfort, not mastery, and that is exactly the right level.
How much will the API calls cost? Most beginners finish this course spending under five dollars in API credits, often zero by using free tiers from Google AI Studio and Hugging Face. Module 5 covers cost control in detail.
Can I use Claude or Gemini instead of OpenAI? Yes. Module 5 covers all three providers and explains tradeoffs. The patterns are the same.
How long does it really take? Most learners finish in 4 to 8 weeks at 6 to 10 hours per week.
Will this course be updated? Yes. The 2026 edition replaces the 2024 stack (older OpenAI SDKs, deprecated models, older LangChain) with current production patterns.
Next step
Start with Module 1: Introduction to AI, ML, Deep Learning, GenAI, and LLMs.