Learn to build the AI systems companies are hiring for.
35 modules, 151 lessons and 20 real projects that take you from Python basics to production RAG, agents and MCP — with interactive labs, module exams, interview prep and a certificate at the end.
- One-time payment, lifetime access
- All future updates included
- Works on Mac, Windows and Linux
Prepare
🎯Interview prep 📖Glossary 🗂️Cheat sheetsYou
🎓Final assessment 🏅CertificateYour learning dashboard
Everything an AI engineer needs — from Python basics to production.
Five phases, 35 modules, roughly 50 hours of reading — laid out in the order a working AI engineer actually needs them. Follow it top to bottom, or jump in at the phase that matches where you are.
Foundations
The ground everything else stands on.
Language models
How the machine actually works.
Retrieval & agents
Where most production work lives.
Production
The part interviews actually probe.
Interview ready
Proving it to someone else.
Pick a track. Earn a certificate that means something.
Three tracks through the same portal, sized for where you are starting from. Your track decides what counts toward your certificate — and the certificate is issued only when every requirement is met.
Actual certificate generated by the portal · downloadable as PNG or PDF
How you earn it
Your profile shows what is still outstanding for your track. Finish the list, sit the final assessment, print the certificate.
- 1Finish the lessons on your trackEvery lesson quiz counted, first attempts only
- 2Pass every module exam70% per exam, and your track's average overall
- 3Clear the exercises and projectsMarked complete in the portal, capstone included
- 4Sit the final assessmentA timed mixed paper from the exam bank
- 🏅Certificate unlockedGenerated with your name, track and date — print or share
Streaks, XP, badges — progress that you can feel.
Five levels from AI Explorer to AI Architect — earn XP for every lesson, lab and project. Keep a streak going. Unlock badges.
XP for lessons, quizzes, labs, exercises and projects. Levels are milestones, not a substitute for the exams.
Nine things you cannot understand by reading about them.
Each lab runs inside the file and is linked from the lesson it belongs to. The Python playground executes real Python in the browser, so you can try every code sample without installing anything.
See how text splits into tokens and what each request costs.
Why did that request cost what it did?
Words placed in a 2D meaning-space you can pan around.
What does "similar meaning" actually mean?
Click a word to see what the model attends to.
What is the model looking at when it predicts?
Watch retrieval, reranking and grounded generation step by step.
Where exactly does a RAG answer come from?
Turn a weak prompt into a production prompt, part by part.
What separates a weak prompt from a production one?
Signal forward, error backward, weights updating live.
What is actually happening during training?
Estimate the monthly API bill before you build the feature.
What will this cost me per month at scale?
An editable table of models you keep current yourself.
Which model should I use for this job?
Run real Python in your browser. No install, no setup.
Can I try the code without installing anything?
Run real Python. In your browser. Right now.
The Python playground executes the code from every lesson without an install, a virtual environment or a terminal. Edit it, break it, run it again.
- Every code sample opens here with one click
- Real Python 3, not a simulation — loops, classes, json, math
- Output streams into the console as it runs
import re, json def tokenize(text): # crude but honest approximation return re.findall(r"\w+|[^\w\s]", text) prompt = "Explain RAG to a backend dev." toks = tokenize(prompt) print(f"tokens: {len(toks)}") print(json.dumps(toks)) print(f"cost @ $2/M: ${len(toks)*2/1e6:.7f}")
Twenty projects you can put on a resume.
Beginner builds to a multi-tenant RAG SaaS. Every project ships with its architecture, reference source code, a run guide and the questions an interviewer will ask about it.
Production RAG SaaS
Multi-tenant, billed, deployed. Users upload documents, ask questions and pay for it — the one project you talk about for the whole interview.
- Tenant isolationRow-level security and per-tenant collections
- Hybrid retrievalBM25 + vectors, reranked before the prompt
- Usage billingMetered tokens, Stripe subscriptions, webhooks
- ShippedDocker, CI/CD and AWS with tracing on
Multi-agent research system
A planner, two researchers and a critic that hand work to each other — traced end to end so you can see why.
MCP server for internal tools
Expose your own tools to any AI client over the Model Context Protocol, with auth.
AI agent with human approval
A tool-calling agent that pauses for sign-off before anything irreversible happens.
refund(order_id="A-2291")From a first chatbot to a voice assistant
Three difficulty levels, each project building on the one before it.
- AI email writer with tone control
- Document summariser (map-reduce)
- AI chatbot with memory
- AI SQL assistant
- Meeting summariser with action items
- Resume analyser and scorer
- Chat with your PDF (RAG)
- AI interview coach
- Semantic search for e-commerce
- Research assistant with web search
- Customer support bot with escalation
- AI coding assistant for a repository
- AI data analyst over CSVs
- Voice AI assistant
- Document intelligence for invoices
- Company knowledge base assistant
Every lesson, built the same ten-step way.
Read it, run it, get it wrong, fix it. Every lesson takes you from a plain-language explanation to working code to a quiz that remembers what you missed.
The RAG pipeline end to end
Retrieval Augmented Generation answers from your documents instead of the model's memory: ingest, chunk, embed, store, retrieve the best chunks for a question, then generate an answer grounded in exactly those chunks.
# the whole pipeline, honestly sizedchunks = chunk(docs, size=500, overlap=80)vectors = embed(chunks) # one call per batchstore.upsert(vectors, metadata=chunks)hits = store.query(embed(q), k=8)hits = rerank(q, hits)[:4] # cheap, big winanswer = llm(SYSTEM, context=hits, question=q)
Chunking by a fixed character count and splitting a table or a code block in half — the retriever then returns fragments the model cannot reason over.
Concept, why it matters and an analogy before any code.
Python and TypeScript examples explained line by line, plus 40 diagrams.
What people get wrong, why, and the takeaways to revise from.
Lesson quiz, module exam at 70%, misses return in the revision queue.
Built to get you through the AI Engineer loop, not just the syllabus.
A dedicated Prepare area sits beside the curriculum: answered questions, cheat sheets, a glossary, troubleshooting for the errors you will actually hit, and a simulator that interviews you.
The four rounds, and what covers each one
- 1Screening
"Explain RAG. What is a token. Why not fine-tune?"
98 answered Q&A · 9 cheat sheets · glossary - 2Technical deep dive
Live coding, debugging a pipeline, reading an eval report.
42 coding exercises · troubleshooting guide · labs - 3AI system design
"Design document QA for 10 million users."
System design module · 4 real-product case studies · 40 diagrams - 4Project walk-through & behavioural
"Walk me through it. What broke? What would you change?"
Interview questions for all 20 projects · behavioural bank · career roadmap
Your own AI tutor, interviewer and project mentor — powered by your API key.
Add an Anthropic or OpenAI key in Profile → AI features and three tools switch on. Everything else works without one — this is a bonus, not a dependency.
Built for four kinds of people.
You need to be comfortable with basic programming; everything else is taught from the start.
Know some Python, never shipped AI. Need a portfolio for placements.
Start: Beginner AI DeveloperBeing asked to "add AI" at work. Skip basics, go to APIs, RAG, agents.
Start: GenAI DeveloperKnow models, need the production side — backends, evals, cost.
Start: AI EngineerInterviews in the next few months. Want one structured place.
Start: GenAI Developer → Interview prepWhat learners say after going through it.
Real feedback from developers and students who bought the portal — on the roadmap, the hands-on sections and how the pieces finally fit together.
“The biggest advantage for me was having everything in one place. Prompt engineering, embeddings, RAG, agents, MCP and AI engineering concepts are explained in a way that's easy to follow.”
“I was confused about where to start with GenAI. This portal gave me a proper roadmap instead of making me jump between random tutorials. The quizzes and practical sections helped me retain what I learned.”
“The AI engineering sections were my favourite. Understanding retrieval, evaluation, cost, latency and observability made me look at GenAI applications much more like a developer and less like just an AI user.”
“Clean interface, structured learning path and lots of interactive content. I especially liked the combination of lessons, quizzes, labs and projects. It feels much more engaging than a normal course.”
“If you already know some programming and want to move into GenAI development, this is a really useful learning resource. The progression from fundamentals to building real AI applications is well thought out.”
“The portal helped me connect the dots between different GenAI concepts. Before this, I knew about embeddings, vector databases, prompting and agents individually. Now I understand how they fit together in an actual AI application.”
“Much more structured than trying to learn GenAI from random YouTube videos and blogs. The roadmap, projects and interview preparation make it useful not just for learning but also for preparing for an AI engineering career.”
Questions people ask before buying.
The short, honest answers to what you get, what you need, and what happens after you buy.
📦What exactly do I receive?
🐍Do I need to know Python already?
🤖Does it cover agents and MCP or just RAG?
🔌Which AI provider do the projects use?
🔑Do I need an API key?
🏅Is the certificate recognised?
💾Will my progress be saved?
🔄Do I get updates?
Start today. Be interview-ready in 16 weeks.
Everything an AI engineer needs in one portal — 35 modules, 20 projects, interview prep and a certificate you actually earn.
- 1Pay once instead of
- 2Get the portal instantlyDelivered the moment payment clears
- 3Open lesson 1Nothing to install. Works on any laptop or mobile.
- 151 lessons
- 9 interactive labs
- 20 projects + source
- 1,363 questions
- 98 interview Q&A
- Certificate
- Free updates for life
- Telegram community ↗