ap.Abhishek Patel

Search

Search notes, essays, projects, the library and the AI course. Use arrow keys to choose and Enter to open.

Free course · Beginner to advanced · Self-paced

Learn AI Engineering from your first word to production.

Never heard of inference, RAG or MCP? Good, that is where this starts. Follow Arjun, a junior developer at a family bakery, as he learns to build real AI products with a patient mentor. Every lesson tells a story, explains one idea in plain words, shows it in an animated clip or diagram, and ends with a quiz. Then you build.

Modules
37
Lessons
199
Projects
17
Hours
~217
0%complete
0/199 lessons 0/17 projects 0 day streak
Start the courseWhat an AI engineer actually doesStart Here

Progress is saved in this browser only.

How this course works

  • A story in every lessonArjun hits a real problem at Crumb & Co., Priya explains it with an analogy, then you see the real thing.
  • Clips, diagrams and videosStep-through animations for the hard ideas, diagrams for every system, and hand-picked videos to go deeper.
  • Quizzes and trackingThree questions after each lesson. Mark lessons complete, keep a streak, and pick up exactly where you stopped.
  • 17 real projectsFrom a support agent to a coding agent and a Perplexity-style search engine, each with a milestone checklist.

The path

All projects →
1Orientation2Foundations3Building with LLMs4Retrieval and RAG5Open models6Agents7Production8Deep learning foundations9Portfolio and career
  1. 1Orientation
  2. 2Foundations
  3. 3Building with LLMs
  4. 4Retrieval and RAG
  5. 5Open models
  6. 6Agents
  7. 7Production
  8. 8Deep learning foundations
  9. 9Portfolio and career

  1. Orientation

    What AI engineering is, the words everyone uses, and the tools on your desk.

    Start hereStart HereYour map to becoming an AI engineer6 lessons3 h1 build+4 added0/6ToolkitDeveloper ToolkitThe tools every AI engineer uses daily5 lessons3 h0/5

  2. Foundations

    Python for AI work and a clear mental model of how language models behave.

    Module 00Python Refresher for AI EngineeringThe Python you need, nothing you don't10 lessons4 h0/10Module 01LLM Mental Model and API SetupHow language models really work9 lessons4 h1 build+1 added0/9

  3. Building with LLMs

    Conversations, prompts, real text tasks, reliable app structure and tool calling.

    Module 02Conversations and Your First ChatbotFrom one question to a real conversation4 lessons2 h1 build0/4Module 03Prompt Engineering: Foundations to AdvancedSay exactly what you mean7 lessons3 h0/7Module 04Real-World Text TasksThe jobs LLMs already do in real products6 lessons14 h2 builds0/6Module 05Structuring End-to-End LLM ApplicationsReliable apps, not lucky demos7 lessons3 h+1 added0/7Module 06Function Calling and Structured OutputsLet the model press buttons5 lessons3 h0/5

  4. Retrieval and RAG

    Embeddings, vector databases and retrieval-augmented generation that scales.

    Module 07Embeddings and Semantic SearchSearch by meaning, not keywords6 lessons11 h1 build0/6Module 08Vector DatabasesStorage built for meaning6 lessons12 h1 build0/6Module 09Retrieval-Augmented Generation (RAG)Give the model an open-book exam8 lessons12 h1 build0/8Module 10Scalable RAG with Async QueuesServing many users without falling over4 lessons2 h0/4

  5. Open models

    Hugging Face models for text, vision, audio and video.

    Module 11Hugging Face: Pre-Trained Models in ProductionThe GitHub of AI models4 lessons2 h0/4Module 12Multi-Modal AI with Hugging FaceImages, audio and video4 lessons2 h0/4

  6. Agents

    Tool-using agents, graphs, memory, voice and the Model Context Protocol.

    Module 13AI Agents: First Agent to Multi-AgentModels that act, not just answer10 lessons5 h+1 added0/10Module 14LangGraph and Stateful Agent WorkflowsAgents as graphs you can debug6 lessons10 h1 build0/6Module 15The Memory Layer in AI AgentsAgents that remember you4 lessons2 h0/4Module 16Graph Memory and Knowledge GraphsMemory with relationships3 lessons85 min0/3Module 17Multi-Modal Agents and Voice AgentsAgents that see, hear and speak4 lessons8 h1 build0/4Module 18Model Context Protocol (MCP) and Agent SDKOne plug for every tool5 lessons11 h2 builds0/5

  7. Production

    Streaming, caching, tracing, evals, security and operating AI for real users.

    Module 19Streaming, Retries, Fallbacks and Rate LimitsAPIs that survive the real world5 lessons2 h0/5Module 20Local LLM DeploymentRun models on your own machine4 lessons2 h0/4Module 21Prompt Caching and Cost/Latency OptimisationFaster and cheaper, same quality4 lessons2 h0/4Module 22Observability and TracingSee inside your AI app4 lessons2 h0/4Module 23Evaluations: The Underrated SkillMeasure quality, don't guess it6 lessons3 h+1 added0/6Module 24Security, Guardrails and Responsible DeploymentSafe by design4 lessons3 h1 build0/4Module 25MLOps and LLMOps in ProductionThe full lifecycle4 lessons2 h0/4Module 26Production Best PracticesThe last mile to reliable6 lessons12 h1 build0/6

  8. Deep learning foundations

    The theory under the hood: PyTorch, CNNs, RNNs, transformers and fine-tuning.

    Module 27Math and PyTorch FoundationsThe engine under the hood7 lessons9 h2 builds0/7Module 28Robust Deep Learning and Computer VisionClean training code and CNNs3 lessons2 h0/3Module 29Sequence Modelling and Multi-IO ArchitecturesModels for order and time4 lessons10 h1 build0/4Module 30Transformers from ScratchThe architecture behind every LLM5 lessons3 h1 build0/5Module 31Deep Generative ModelsHow machines learn to create3 lessons2 h0/3Module 32LLM InternalsTheory to match the practice5 lessons3 h0/5Module 33Fine-Tuning, LoRA/QLoRA and Trending TopicsTeach a model new habits4 lessons3 h1 build0/4

  9. Portfolio and career

    Five portfolio projects that pull everything together, and getting hired.

    FinalePortfolio Projects and CareerBuild the things that get you hired8 lessons42 h5 builds+3 added0/8

What you will build

Project details →
  1. Module 04Working with the OpenAI API
  2. Module 04Customer support agent
  3. Module 07Exploring embeddings
  4. Module 08RAG chatbot over YouTube transcripts
  5. Module 09Enterprise advanced RAG
  6. Module 14A Claude Code-like coding agent
  7. Module 17Voice-based AI agent
  8. Module 18AI code reviewer like CodeRabbit
  9. Module 26AI trading research agent
  10. Module 27a neural network in TensorFlow
  11. Module 27MNIST autoencoder
  12. Module 29multi-input models for OCR
  13. FinaleChatGPT-style chatbot platform
  14. FinaleChat with any document
  15. FinaleAI search engine like Perplexity
  16. FinaleAI web app builder like Lovable
  17. FinaleAI interviewer like Mercor

Never heard these words? Start here.

The whole glossary →
Large language model
A very large model trained on huge amounts of text to predict the next piece of text.
Token
A small chunk of text, often a short word or part of a longer one, that a model reads and writes.
Inference
Using a trained model to get an answer, with its parameters fixed.
Prompt
All the text you send a model as input: instructions, information and the question.
Embedding
A list of numbers that captures what a piece of text means, so similar meanings get similar numbers.
Retrieval-augmented generation
Finding relevant passages from your own documents first, then giving them to the model so it answers from real facts.
Agent
A model running in a loop that decides its own next step, uses tools, checks results and keeps going until a task is done.
Model Context Protocol
An open standard for connecting AI apps to tools and data sources in one common way.
Context window
The maximum amount of text, counted in tokens, that a model can consider at once.
Hallucination
When a model gives a fluent, confident answer that is false or made up.
Fine-tuning
Training an existing model a little more on your own examples so it picks up a style, format or narrow skill.
Eval
A repeatable test that measures how good an AI system's answers are across many examples.
Begin with lesson one Browse the glossary