SPYRIDON KTENAS

SPYRIDON KTENAS


Research Focus & Metrics

Academic Career Length

1Year (2026-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
9 November 2026 in Preprints.org
Verified
Preprint
Open access
PDF file
SKSPYRIDON KTENAS
SPYRIDON KTENAS
MGMirto M. Gasparinatou
Mirto M. Gasparinatou
Panagiotis VlamosPanagiotis Vlamos
Panagiotis Vlamos
Large language models and coding agents are changing how digital solutions are specified, built and maintained. Conventional software lifecycles, however, assume extended manual effort, relatively stable requirements and sequential progression, and they offer little protection against non-deterministic generation, self-confirming validation and nominal human review. Following Design Science Research, this study designed IntDEx (Intelligence-native Delivery Experience), an engineering framework for delivering digital solutions with AI as the primary enabling layer, and evaluated it in a single real case: the construction of a sovereign data-science platform for public-health experimentation. IntDEx combines a seven-phase non-linear delivery flow, an accountability-based role model and graded conformance profiles with two assurance constructs, enforceable human oversight and evidential independence. A solution of 52,598 lines of code was rebuilt by a model from a different provider using IntDEx artefacts alone, reproducing approximately 95% of the tested routes and user journeys, whereas an intent-only baseline produced a non-working system. Security controls that were implicit in the original implementation were weakened in the rebuild. Enforced checkpoints changed the outcome in 7 of 17 major decisions, and cross-provider review led to 4 human-confirmed corrections that self-assessment would not have produced. These results indicate feasibility rather than general effectiveness, and multi-case replication is required.