*Changes may apply
Module 1: Fundamentals of Generative AI
- Introduction to GenAI: History, evolution, and core concepts (Generative vs. Discriminative AI).
- Market Landscape: Overview of leading tools and platforms.
- LLM Architecture: Deep dive into the Transformer architecture, tokenization, and embeddings.
- Model Types: Understanding the differences between open-weight and closed-weight models.
Module 2: Prompt/Context Engineering & Text Applications
- Working with LLMs APIs: including Google Gemini, OpenAI, Anthropic as well as open-source models from HuggingFace.
- Prompt/Context Engineering Principles: From zero-shot and few-shot to advanced prompting techniques like chain of thought (CoT) and context handling.
- Understanding the Parameters: all the settings you can tweak to get the most of your LLMs.
- Evaluation driven development: how to evaluate GenAI products and what is the correct process?
- Reasoning: The evolution of reasoning, from simple CoT to today’s reasoning models
- Company Brain: large scale context management
Module 3: Retrieval Augmented Generation (RAG)
- RAG fundamentals: The need for external memory, vector databases, and semantic search.
- Basic pipelines: Building a standard RAG system.
- Advanced RAG: Re-ranking, hybrid Search, and handling complex documents.
- Evaluation of RAG systems: Metrics and methods for evaluating RAG performance.
Module 4: Fine-tuning
- Fine-tuning: what does it mean?
- Parameter efficient fine-tuning: including LoRA.
- Peeking inside the model: An introduction to mechanistic interpretability, understanding how a model works internally, from features and circuits to what individual neurons actually represent.
- Understanding fine-tuning process of reasoning models: We will cover the underlying theory and conceptual foundations, without engaging in large-scale fine-tuning in practice.
Module 5: AI Systems
- Workflows & Systems: what is an AI system and why RAG is just a specific case of it? Reviewing common architectures.
- Automation tools: such as n8n.
- ADLC: Coding better with AI – Cursor, Claude, TDD, rules, and a full development harness
- LLM cost management: offline and online
- Local model hosting
- Harness engineering: Skills, rules, model routing, real-time LLM decision making
Module 6: AI Agents
- Development AI Frameworks: Introduction to LangChain, LangGraph and LlamaIndex.
- Tool use: Function calling and connecting LLMs to external APIs.
- MCP & other protocols (such as A2A): Tool use 2.0.
- AI agents: Differences between chatbots and agents (e.g., ReAct design pattern).
- Introduction to Google ADK.
- Agents evaluation.
Module 7: Final Project
- Ideation & Planning: Selecting a business problem and defining the solution architecture.
- Development: Hands-on implementation of a GenAI-based product.
- Demo Day: Final presentation of the projects.