Marko Chalupa
CTO, AI Architect, and founder of SnapOS Foundation. Over 20 years building production systems that make consequential decisions at scale. The governance gap Decision Integrity addresses is not a theoretical observation — it comes from watching real systems continue acting after the basis for their decisions had already changed.
Why this exists
Decision Integrity did not emerge from a research paper. It emerged from 20+ years of building and operating production software — trading systems, healthcare platforms, e-commerce infrastructure, governance tools — for organizations where automated decisions have real financial, regulatory, and operational consequences.
The pattern that kept appearing was not that systems broke. It was that systems continued — operating under authorizations, assumptions, and mandates that had quietly changed. Monitoring was green. Controls were documented. Compliance was maintained. And still, the decision being executed was no longer the decision that had been approved.
No existing governance framework had a name for this. No standard addressed it directly. That gap is what SnapOS is built to close.
Background
Marko Chalupa is CTO and AI Architect at United Coding GmbH & Co. KG, based in Mühltal near Darmstadt, Germany. He has been self-employed for over 20 years, with delivery experience across web, software, and application development for clients ranging from DAX-listed companies to public institutions.
In parallel, he founded IT-Schnittstelle GmbH, a governance audit practice focused on decision architecture and structural accountability in German-speaking organizations.
SnapOS Foundation is the research and specification layer above both — where the formal frameworks, governance protocols, and category definitions are developed, published, and maintained. The combination of operative delivery history and formal specification work is what makes the claim credible: this is not governance theory from someone who has never shipped a production system.
What qualifies this work
Decision Integrity is not a consulting methodology rebranded as research. The frameworks are formally specified, timestamped, and published under citable DOIs.
Capability and compliance evaluation framework for AI systems under drift. Published on Zenodo. Submitted to the EU AI Office Expert Forum (Contribution ID: 510c3274). DOI: 10.5281/zenodo.18362037
30-day operational audit protocol for mandate drift detection in high-stakes AI systems. DOI: 10.5281/zenodo.18824037
Governance layer for continuous decision legitimacy through Authority Closure, Assumption Closure, and Evidence Closure. zenodo.org/records/19135942
Canonical field definition for SSE as the scientific foundation underlying Decision Integrity analysis. Published November 2025. DOI: 10.5281/zenodo.17711427
Black-box diagnostic framework for multi-turn instability in language models. zenodo.org/records/18361523
Formal submission positioning European AI governance infrastructure as a strategic priority. Contribution ID: 510c3274. GCCL v0.1 submitted as candidate framework for AI capability and compliance assessment.
Persistent research identity linking all SnapOS publications. orcid.org/0009-0000-6493-4599
Three layers — one coherent system
snapos.org
Formal specification, frameworks, publications, and category definitions for Decision Integrity and Semantic Stability Engineering.
united-coding.com
Software, web, and application development. 20+ years of production delivery. The operative foundation that makes framework claims credible.
it-schnittstelle.de
German-language governance audit practice for organizations navigating decision architecture, structural accountability, and re-legitimation.
What this is — and what it is not
SnapOS Foundation presents itself as an independent research and governance initiative. It does not claim formal certification authority, official EU institutional status, or regulatory mandate unless explicitly stated.
The frameworks presented here — Decision Integrity, DASR, DCF, GCCL, DIP — are research, governance, and specification materials. They are published under citable DOIs, submitted to relevant expert forums, and designed to be operationally applicable. They are not yet ratified standards, and that distinction is stated clearly.
The goal is to define a governance category that does not yet formally exist in standards bodies — and to do so with enough rigor, specificity, and documented evidence that the category becomes recognizable, referenceable, and eventually adoptable.