Get in touch
← All courses

AI Engineering Fundamentals

or what AI is all about

A hands-on introduction to modern AI systems: from neural networks and the Transformer architecture through LLM programming, AI agents and RAG, all the way to AI-assisted software development.

Get in touch Syllabus

The world of AI keeps changing at a record pace: new frontier models (GPT, Claude, Gemini), the MCP standard and skills, the widespread adoption of agent-based systems and a whole new generation of AI-powered development tools are reshaping the landscape right now.

AI Engineering Fundamentals is a crutch for catching up and getting the fundamentals down fast. Building on the timeless foundations of the previous edition, it returns with an updated model overview, a reworked agents module and current topics, so you get not just the basics, but the knowledge that actually matters today.

The goal of the course is to show, from a practical angle, how modern AI systems are built and how they work, how large language models (LLMs) "think", and how to apply them effectively to real-world problems. Step by step, participants learn the knowledge and skills needed to build AI components, LLM-based applications and even their own AI agents on their own, while getting a clear view of the entire AI ecosystem.

8 modules

Syllabus

  1. Deep learning fundamentals

    • The goal of AI and the machine learning workflow
    • From biological to artificial neurons (weights, bias, activation functions)
    • What makes a network "deep"? Understanding layers
    • How networks learn: loss functions, gradient descent, backpropagation
    • What are hyperparameters? (e.g. learning rate, batch size)

  2. The LLM revolution and the Transformer architecture

    • What is a large language model (LLM)?
    • The problem with traditional neural networks on sequential data
    • How LLMs understand text: tokens and embeddings
    • The heart of LLMs: the Transformer architecture
    • The magic ingredient: the self-attention mechanism (Q, K, V)
    • The LLM training lifecycle: pre-training, SFT, RLHF, RLVR
    • The probabilistic nature of LLMs and why they "hallucinate"

  3. The AI ecosystem today

    • What is a reasoning model? How do today's frontier AIs "think"?
    • Chain-of-thought and extended thinking: when is more thinking time worth it?
    • Prompting basics: zero-shot, few-shot, CoT, role-based techniques
    • Frontier vs. open-weight models: GPT, Claude, Gemini and their open-weight challengers
    • MoE architectures and efficient inference
    • Multimodal models: text, image, audio, video
    • Running models locally: HuggingFace, Ollama, LM Studio

  4. Programming LLMs: APIs and frameworks

    • Accessing LLMs via APIs: OpenAI, Google Gemini, Anthropic Claude
    • Key API parameters: model, messages, temperature, reasoning effort
    • Generating structured outputs with Pydantic AI

  5. Building AI agents

    • What is an AI agent? (LLM + memory + tools + planning)
    • The reasoning loop: ReAct, Reflection and Planning patterns
    • Tool use: functions, API calls, search
    • Model Context Protocol (MCP): the de facto standard for connecting LLMs to external systems
    • Skills: a simple way to extend what agents can do

  6. Extending knowledge – RAG and fine-tuning

    • The core problem: stale data and missing domain-specific knowledge
    • Strategies: fine-tuning vs. RAG vs. long context
    • The RAG pipeline: ingest → embed → store → retrieve → generate
    • Embedding models and semantic search
    • Vector databases (Chroma, Qdrant, pgvector)
    • Agentic RAG: when do you outgrow the classic pipeline?

  7. Evaluating and securing AI systems

    • How do we know our AI is any good? Evaluation metrics and benchmarks
    • LLM-as-a-Judge: automating qualitative evaluation
    • Prompt injection: how LLM-based systems get attacked
    • Secure system design: guardrails, input/output filtering
    • Responsible AI basics: bias, fairness, transparency

  8. AI-assisted development today

    • The limits of chat interfaces for coding: why ChatGPT or Claude.ai isn't enough
    • IDE integrations: Cursor, Copilot, Cline, Kilo Code – what's the difference?
    • CLI-based development: Claude Code, Codex – what's different about AI development from the terminal?
    • Vibe coding vs. agentic engineering
    • Multi-agent coding workflows: parallel agents, code review, automated testing

Interested in this course?

Drop us a line and we'll figure out how to tailor the course to your team.

Get in touch