AI & Machine Learning Roadmap: From Basics to LLMs

Learn by doing. Build by understanding. Master by creating.

Open-source AI education built by a student, for students and learners worldwide.

Basic to Expert. Zero to Language Models to AI Agents. 82 lessons. 100% hands-on.

Quick Start • Student Guide • Exam Prep • LinkedIn • Support Project

This is an independent learning project, NOT official University of Auckland, material. Use responsibly and follow your institution's academic integrity policies. See Academic Integrity Policy for details.

A structured, hands-on learning path from basic arithmetic to complete language models and intelligent AI agents. 82 lessons with runnable code, visualizations, and practical projects.

Perfect for:

• University students learning AI/ML
• Self-learners building AI skills
• Professionals upskilling in deep learning
• Anyone wanting to understand AI from first principles

Build intelligent agents that reason, plan, and act! The new Agents sector teaches you how to create production-ready AI agents using industry-standard frameworks.

What You'll Build:

• Conversational agents with memory
• RAG systems for knowledge retrieval
• Multi-agent systems with specialized roles
• Production APIs and user interfaces
• Real applications: code assistants, data analysts, research agents

18 comprehensive lessons | 4-6 hours of complete content ready now | Start Learning - Agents Sector

# Create virtual environment
python -m venv .venv
source .venv/bin/activate  # macOS/Linux
.venv\Scripts\activate     # Windows

# Install dependencies
pip install tensorflow torch numpy matplotlib jupyter

jupyter lab
# Open any notebook from Basic/ folder

Or use Google Colab (no setup needed) - Click "Open in Colab" badge in any notebook.

graph LR
    A["Basic<br/>(20 lessons)<br/>Foundations"] --> B["Intermediate<br/>(15 lessons)<br/>Advanced ML"]
    B --> C["Advanced<br/>(15 lessons)<br/>Production ML"]
    C --> D["Expert<br/>(15 lessons)<br/>Research"]
    
    B --> E["Agents - NEW<br/>(18 lessons)<br/>LangChain/LangGraph"]
    C --> E
    
    style A fill:#ffffff,stroke:#000000,stroke-width:2px,color:#000000
    style B fill:#f0f0f0,stroke:#000000,stroke-width:2px,color:#000000
    style C fill:#e0e0e0,stroke:#000000,stroke-width:2px,color:#000000
    style D fill:#d0d0d0,stroke:#000000,stroke-width:2px,color:#000000
    style E fill:#c0c0c0,stroke:#000000,stroke-width:3px,color:#000000

Learning Paths:

• Traditional ML/DL: Basic to Intermediate to Advanced to Expert
• Agent Development: Basic to Intermediate to Agents (NEW)
• Full Stack AI: Basic to Intermediate to Advanced to Agents to Expert

AI/
├── Basic/              # 20 Lessons (B01-B15 + B01a, B05a, B05b, B05c, B09a, B09b, B10a) [COMPLETE]
├── Intermediate/       # 15 Lessons (I01-I15) [COMPLETE]
├── Advanced/           # 15 Lessons (A01-A15) [COMPLETE]
├── Expert/             # 15 Lessons (E01-E15) [COMPLETE]
├── Agents/             # 18 Lessons (AG01-AG18) - LangChain & LangGraph [NEW]
├── application/        # Live demos & practical implementations
│   ├── compsci713/     # COMPSCI 713 weekly apps (Wumpus, KG, RNN, NEAT, Q-Learning)
│   └── compsci714/     # COMPSCI 714 weekly apps (Gradient Descent, CNN, Transformer, BPE)
├── documentation/      # Guides & resources
└── landingpage/        # Landing page assets

Interactive demonstrations of AI concepts in action:

Interactive AI demos — playable in your browser

View All Games • Explore the application/ folder for source code and deployment guides.

COMPSCI 713 Complete Guide · COMPSCI 714 Complete Guide · COMPSCI 769 Complete Guide

• Before lectures: Review relevant Basic lessons
• During semester: Build practical projects from examples
• For assignments: Use as reference, implement your own
• For exams: Review all concepts in relevant lessons

Complete Student Guide

• Symbolic logic & first-order logic
• Tensors & linear algebra
• Linear regression & gradient descent
• Binary & multi-class classification
• Neural networks from scratch
• Training & optimization theory (COMPSCI 714)
• Data preprocessing & evaluation
• Regularization & overfitting
• CNNs, RNNs, Transformers
• Tokenization & language models

• Advanced optimization & regularization
• Transfer learning & domain adaptation
• Object detection & segmentation
• Seq2seq & advanced transformers
• Hyperparameter tuning & AutoML
• Generative models (VAEs, GANs)
• MLOps & deployment

• Fine-tuning LLMs
• Prompt engineering & RAG
• Vision-language models
• Distributed training
• Mixed precision & inference optimization
• ML pipelines & monitoring
• Responsible AI

• Reading & implementing research papers
• Neural architecture search
• Meta-learning & few-shot learning
• Deep reinforcement learning
• RLHF & alignment
• Federated learning
• Cutting-edge research

• AI agent fundamentals & ReAct pattern
• LangChain basics: chains, prompts, parsers
• Memory systems (buffer, summary, vector)
• Tools and function calling
• Building RAG agents
• LangGraph state machines
• Multi-agent systems
• Human-in-the-loop patterns
• Production deployment (APIs, UIs, monitoring)
• Real-world applications: code assistants, data analysts, research agents

Full Agents Curriculum | Getting Started Guide

Current Status: Foundation Complete (AG01-AG03) | View Roadmap

Next Up: AG04 (Tools & Function Calling) then AG05 (First Agent) then AG06 (RAG Agents)

• Q-Learning & reinforcement learning (grid world)
• NeuroEvolution (NEAT) for topology search
• Interactive model evaluation dashboards
• RNN/LSTM/GRU sequence modelling
• CNN feature map visualization
• Transformer attention visualization
• BPE tokenization exploration

Beginner: Sentiment analysis, image classifier, text generator, spam detector, digit recognition

Intermediate: Medical image analysis, chatbot, stock predictor, document summarizer, multi-label classification

Advanced: RAG system, domain-specific LLM, multi-modal search, code reviewer, real-time detection

Agents: Personal research assistant, automated code review agent, customer support system, data analysis agent, content creation pipeline

Research: Novel architecture, paper reproduction, bias detection, model compression, federated learning

Appropriate Use:

• Learning concepts and understanding implementations
• Preparing for lectures and exams
• Using as inspiration for original projects
• Understanding different approaches

❌ Inappropriate Use:

• Copying code for assignments without understanding
• Submitting repository code as your own work
• Using during closed-book assessments
• Violating your institution's policies

Full Academic Integrity Policy

Contributions welcome! See Contributing Guide for details.

• Questions? Check Documentation Index
• Issues? Open a GitHub issue
• Suggestions? Submit a pull request
• Connect: LinkedIn
• Discord: Join the discussion

Created by a student pursuing a Master of Artificial Intelligence at the University of Auckland.

Why this exists: To make quality AI education accessible to everyone, combining theory with practical implementations.

MIT License - See LICENSE for details.