AI Consultant · Former CTO · Prague & remote
I spent nine years as a CTO, most recently turning an engineering team AI-native and building AI agents that run in production. I help software teams turn AI experiments into reliable workflows, with the tools and skills to keep going.
Why companies call me
Pilots that never ship
Impressive demos that die in front of real data, security and edge cases.
Faster typing, same delivery
Developers use AI, yet features still take weeks, because the process didn't change.
No one to own it
Leadership knows AI matters but lacks someone who's done it end to end.
Selected AI work
Engineering teams, production agents and hands-on workshops. Here’s what I built, what changed, and what I learned.
How a ~20-person engineering team moved from writing code by hand to directing AI agents, with product managers and designers now working directly in the codebase.
A regression suite of over 3,000 tests kept alive by AI agents. Agents tell flaky or outdated tests apart from real product bugs, repair the tests, and file evidence-backed bug reports.
A full-day workshop for a team building desktop software that drives physical hardware. We started with an AI-maturity assessment and finished with working agent tooling in their own repository.
A multi-agent system that takes a feature from stakeholder request through research, prototype, plan, code, review, QA and release, with a human approving every step.
Stakeholders describe a problem in a Teams channel. An AI intake agent checks whether it's already planned or shipped, asks questions, and files a structured request that a domain expert triages. The outcome goes back to the same thread.
An internal planning tool that uses LLMs and semantic search over past work to estimate effort, predict delivery delays and surface productivity issues for the leadership team.
Agents can only close the loop when they can run, seed and observe the system on their own. I built the test data, production-like data and payment simulation that made that possible.
Services
From the first conversation about AI to agents running in production and a team that can carry on without me.
An honest assessment of where AI creates value in your business and where it doesn't. Includes an AI-maturity baseline, prioritised use cases with a business case, and a 90-day roadmap you can actually execute.
Assessment · Use-case map · Roadmap
Help your team delegate work to AI agents, review the results and ship with confidence. Hands-on coaching in your own codebase, with shared context, clear specifications and automated QA.
Workshops · Team coaching · Repo setup
Connect AI agents to your tools and workflows. I build integrations, approval steps, monitoring and cost controls, then help your team run and maintain them.
Prototype · Production build · Handover
Senior technology leadership a few days a month. Get support with architecture, hiring, vendor choices and board reporting, with clear priorities for your product and team.
Retainer · Board reporting · Hiring
The lab
Small personal projects are where I test new models, frameworks and patterns before recommending them to clients.
Natural-language food tracker
Free-text meals ("1 beer, schnitzel with mash") parsed by an LLM into nutrition data, enriched from Open Food Facts.
Next.js · Claude · Supabase
Household expense categoriser
Parses bank statements from three banks. Rules first, LLM second, and the rule set grows from approved results.
React · Claude · Google Sheets
LifeCrit
A family self-improvement RPG with AI-generated quests through structured LLM output.
Next.js · Vercel AI SDK
Spanish flashcards for kids
Text-to-speech, quizzes and AI-generated practice sentences built from the child's own words.
Vercel · OpenRouter
CTO operating system
Runbooks, OKRs, 1:1s and decision logs in Markdown, with Claude as a daily sparring partner.
Claude Code
AI candidate triage
Pulls applications from the recruiting system and scores CVs with structured output for a faster first pass.
Claude · JSON schema
How we'd work together
We talk about your goal. You get my honest view on whether and how AI helps.
A short, fixed-price assessment of your processes, data and team, ending in a concrete plan.
We ship the first use case together and your team learns by doing, not by watching slides.
Documentation, guardrails and skills stay with you. No lock-in, no black boxes.
About
I've built software products for over 15 years: from a web developer in a Prague agency, to the second engineer at HotelQuickly in Bangkok, to founding CTO of GuestReady (London, $8M+ raised, 5 acquisitions integrated) and then CTO of Vicoland in Frankfurt, starting in 2020.
At Vicoland I led the move to AI-native development: a team that directs AI agents and reviews their work, AI that runs QA every night, and agents that move features from request to release. One lesson I bring to client work: keep the system simple.
I also co-founded Naučmese.cz, the largest Czech marketplace for courses (55,000+ students), so I think about AI in terms of business outcomes, not technology for its own sake.
FAQ
You tell me what you want to achieve, and I ask a few questions about your team, data and processes. You leave with an honest opinion and a concrete next step. If I'm not the right fit, I'll say so.
Mostly software companies and tech-enabled businesses with roughly 10 to 300 people, where one experienced person can make a visible difference quickly.
Both. I've been a hands-on engineer for 15+ years and still write code every day. Most engagements combine strategy with a first production use case.
Whatever fits the problem. Mostly Claude and Claude Code, OpenAI and open-weight models through OpenRouter, MCP for integrations, and vector search where retrieval matters. I'm vendor-neutral.
Yes. I've led fully remote teams across Europe and Asia for over a decade. I work in English and Czech, remotely or on-site for workshops.
Let's talk
Tell me where your team is stuck. In a free 15-minute call, we'll discuss what's realistic, where to start, and whether I can help.
I reply within one working day.