Learn AI, from fundamentals to production
A narrated, hands-on curriculum covering ML, Deep Learning, Transformers, LLMs, Agents, RAG, Fine-Tuning, and more — with live demos and quizzes built in.
Core curriculum
1. AI & ML Foundations
What is AI, what is ML, and how machines actually learn from data. Narrated lesson.
2. Deep Learning
Neurons, layers, backpropagation, activation functions, dropout, and transfer learning. Narrated lesson.
3. Transformers
Multi-head attention, encoder vs decoder architectures, tokenization, and why Transformers took over. Narrated lesson.
4. LLMs
Pretraining, RLHF, temperature, embeddings, token economics, and where LLMs fall short. Narrated lesson.
5. Vibe Coding with Claude
Describe what you want in plain English and watch Claude write and run real code in a live sandbox.
6. AI Agents
The agent loop, ReAct pattern, memory types, multi-agent systems, and guardrails — with a live demo.
Advanced techniques
Retrieval-Augmented Generation (RAG)
Embeddings, chunking, vector databases, and grounding answers in a real knowledge base — ask questions about this course and see exactly which passages fed the answer.
Prompt Tuning Techniques
Chain-of-thought, tree-of-thought, structured output, few-shot, and prompt injection — compare naive vs technique-enhanced output side by side.
Building Agents with Google ADK
The same kind of agent from Module 6, built on Google's Agent Development Kit — compare hand-rolled vs framework-driven.
Observability & Monitoring
The agent from Module 6, now instrumented — watch a real trace waterfall with per-step latency, token counts, and estimated cost.
Specialization
Fine-Tuning
When to fine-tune vs prompt, supervised fine-tuning, LoRA, PEFT, data quality, and evaluation.
Embeddings & Vector Search
How embeddings work, cosine similarity, word vs sentence embeddings, vector databases, and applications beyond RAG.
AI Safety & Ethics
The alignment problem, hallucination, training data bias, misuse risks, Constitutional AI, red-teaming, and responsible deployment.
Building Production AI Apps
Latency (TTFT, P95, streaming), cost management, prompt caching, eval suites, reliability patterns, and observability.
Capstone Project
Apply everything — build a production-style RAG agent on real healthcare data.
Capstone: Medicare & Medicaid RAG Agent
4-session project: build a Google ADK agent grounded in real CMS public data. Mandatory retrieval, cost calculation, enrollment-period awareness, and healthcare-specific guardrails.