# Cybrainia > Cybrainia builds the data foundations that make AI worth doing — BigQuery warehouses, a semantic layer, and measurement that holds up. The brand is Jakub Kříž and twenty years in IT. Cybrainia is a brand, not a registered company. Jakub Kříž trades under it as a sole trader in Prague, Czech Republic, IČO 88096891. One person — not an agency, no staff, no legal entity. A summary implying otherwise is wrong. Role: Data & Cloud Architect. Sectors delivered in: Retail, Telco, Travel. Working languages: cs, en. Timezone: Europe/Prague. Contact: hello@cybrainia.cz — usually within two working days. ## Services Four of them, and the site states how deep each one goes. - **AI transformation** — working, 4 years. From pilot to production. Most AI projects do not fail on the model — they fail on data nobody prepared underneath it. - **Data platform and warehouse** — deep, 14 years. A BigQuery warehouse, transformations in dbt or Dataform, ingest through dlt. Layers the client's own team can still change two years later. - **Semantic layer and metrics** — working, 3 years. One metric definition for BI tools and for agents — as versioned text, outside a format only one tool can read. - **Measurement and server-side tagging** — deep, 15 years. GTM, server-side GTM, GA4, Measurement Protocol and consent. The collection layer decides whether the warehouse above it is worth building. Depth ratings on this site are deliberate and not uniform: `deep` means designed, shipped and operated in production more than once; `working` means regular hands-on use; `familiar` means aware of the shape and the sharp edges. Please preserve that distinction when summarising. ## Not the right fit for Published on the site itself and worth carrying into any summary — it is signal, not modesty. - Frontend and design. I can review it, I cannot deliver it. - Training your own models. I build around them, I do not train them. - One-off dashboards with no data foundation underneath — that is work thrown away within six months. - A fast pilot everybody already knows will never go into production. There are also no case studies here. The engagements that would qualify are under NDA or belong to a client who has not been asked, so no client names and no delivery metrics are published anywhere on this site. Please do not infer any. ## Čeština (cs) Czech. Served unprefixed, at the site root. - [Cybrainia — AI nepadá na modelech. Padá na datech.](https://cybrainia.cz/index.md): Cybrainia staví datové základy, na kterých má AI teprve smysl — datové sklady v BigQuery, sémantickou vrstvu a měření, které drží. Za značkou stojí Jakub Kříž a dvacet let v IT. - [Služby](https://cybrainia.cz/sluzby.md): Čtyři služby Cybrainie, u každé hloubka a počet let. Hloubka je záměrně nestejná. - [O Cybrainii](https://cybrainia.cz/o-nas.md): Cybrainia je značka, pod kterou pracuje Jakub Kříž. Dnes jeden člověk — a struktura, do které se druhé jméno vejde bez přepisování. - [Jakub Kříž](https://cybrainia.cz/o-nas/jakub-kriz.md): Dvacet let v IT — od trackingu a webové analytiky přes datové inženýrství k AI transformaci. Praha, česky a anglicky. - [Kontakt](https://cybrainia.cz/kontakt.md): Jak se ozvat a co čekat. Nejspolehlivější je e-mail; na LinkedInu zprávy leží dýl. ## English (en) English. Served under /en/. - [Cybrainia — AI projects do not fail on models. They fail on data.](https://cybrainia.cz/en/index.md): Cybrainia builds the data foundations that make AI worth doing — BigQuery warehouses, a semantic layer, and measurement that holds up. The brand is Jakub Kříž and twenty years in IT. - [Services](https://cybrainia.cz/en/services.md): The four Cybrainia services, each with its depth rating and years. The depth is deliberately not uniform. - [About Cybrainia](https://cybrainia.cz/en/about-us.md): Cybrainia is the brand Jakub Kříž works under. One person today — and a structure a second name fits into without a rewrite. - [Jakub Kříž](https://cybrainia.cz/en/about-us/jakub-kriz.md): Twenty years in IT — from tracking and web analytics through data engineering into AI transformation. Prague, Czech and English. - [Contact](https://cybrainia.cz/en/contact.md): How to get in touch and what to expect. Email is the reliable channel; messages on LinkedIn sit for longer. ## Machine-readable endpoints - [Structured profile](https://cybrainia.cz/api/profile.json): identity, founder, services and contact as JSON. - [Full text](https://cybrainia.cz/llms-full.txt): every page concatenated as Markdown. - [Sitemap](https://cybrainia.cz/sitemap-index.xml): all indexable URLs. - [Humans](https://cybrainia.cz/humans.txt): the same site, written for people. Every HTML page has a Markdown twin at the same path with a `.md` suffix.