Courses
Vibe Coding with AI Agents
“Vibe coding” — describing what you want and letting an AI agent write the code — is fast, but only useful if you can steer it and catch its mistakes. This hands-on workshop shows how to build working software with Claude Code, GitHub Copilot, Antigravity, and Codex while keeping the result maintainable and secure. Course Content How AI coding agents actually work: context windows, tools, and where they go wrong Writing prompts and specs that produce correct, reviewable code the first time Reviewing and refactoring AI-generated code — spotting the typical failure patterns Setting guardrails: tests, linters, and git workflow so an agent can’t wreck your project Building a small feature end to end, from prompt to committed, tested code Formats & Pricing The same course, in three formats — pick what suits you:
GenAI Integrations
Learn how to connect generative AI to real systems: retrieval-augmented generation over your own data, custom MCP servers that give agents safe access to your tools, and autonomous agents orchestrated to do useful work. Course Content Architecture of GenAI integrations: models, embeddings, retrieval, and orchestration Building a RAG pipeline over your own documents (vector database, chunking, retrieval) How the Model Context Protocol (MCP) works and building your own MCP server (FastAPI / Node.
Drupal Basics
Drupal has a reputation for a steep learning curve. This beginner course gets you productive fast by building a real site, explaining Drupal’s mental model as you go, and using an AI agent to shorten the path — while making sure you understand what it produces. Course Content Getting a local environment running with DDEV and Docker How Drupal thinks: entities, fields, content types, and configuration Building pages and site structure without code (blocks, menus, views) Theming basics: templates, regions, and Twig Your first custom module — scaffolded with an AI agent and understood line by line Requirements: basic command line and a laptop with Docker (setup instructions sent in advance).
Drupal Pro Developer
Advanced Drupal architecture for developers who already know the basics. Build custom modules the right way, integrate external services, and deploy a modern RAG chatbot into the Drupal admin — with an AI agent accelerating the work and a code review discipline to keep it correct. Course Content Structuring custom modules for Drupal 10/11 — dependency injection, services, event subscribers Reviewing and fixing the mistakes AI agents typically make in Drupal code (cache, routing, permissions) Connecting to external databases and vector stores (Qdrant, ChromaDB) Deploying a RAG chatbot and integrating it into the Drupal editorial system Setting up a CI/CD pipeline (GitHub Actions) to deploy automatically — demo plus a reusable template Requirements: PHP fundamentals or solid experience with AI coding agents, plus basic git and Docker.
Practical OSINT & Threat Intel
How to collect, analyze, and safely evaluate open-source intelligence (LinkedIn, search engines, file metadata) and feed it into intelligence RAG bots. Course Content Collecting information from open sources without leaving footprints Scraping and structuring data from LinkedIn and social media Extracting and analyzing file metadata (EXIF, PDF metadata, etc.) Security when conducting research (OPSEC, shadow profiles) Formats & Pricing The same course, in three formats — pick what suits you:
SEO Tools in Practice
Run a real on-page SEO analyzer end to end: crawl a site, turn the results into actionable audits, and let AI agents propose fixes. Based on a production open-source tool built on Crawlee and Playwright, with an MCP server that exposes the whole workflow to AI agents. Course Content Running the analyzer with Docker Compose — the crawl-once, run-reports workflow What the crawler extracts: titles, meta descriptions, canonicals, hreflang, structured data (JSON-LD), broken links Reading the Markdown audit and turning issues into a prioritized developer backlog LLM-powered title and meta-description fix suggestions — and when to trust them Exposing the analyzer to AI agents via its MCP server for automated audits Formats & Pricing The same course, in three formats — pick what suits you:
Peripateo — Walking Mentorship
Peripateo is a different kind of session: no room, no slides. We walk through Prague and talk through whatever you’re stuck on — a software architecture decision, an AI integration you’re planning, a codebase you’ve inherited, or a career crossroads. Walking loosens thinking, and one-on-one means the whole session is about your question. What to bring A concrete problem or decision you want to work through Any context you can share in advance (repo, diagram, a few sentences) so we start fast Comfortable shoes How it works In-person only, in Prague — we agree a route and a starting point One-on-one, subject entirely up to you: architecture, AI strategy, code review, career, or a plain sounding board You leave with clear next steps, not just a conversation Duration: per session (2–3 hours)