{{ r.range }}
{{ r.duration }}
- {{ b.k }}{{ b.t }}
AI Engineer. .NET specialist.
I build and maintain enterprise .NET systems — banking, payments and fintech data. Since 2024 I have also been changing how I build them: agentic coding workflows, structured LLM outputs for data conversion and code quality, and ABSD, an open-source multi-agent method for software development that I designed. At EY I also counsel junior developers and lead the team's day-to-day work.
Nicosia, Cyprus · {{ years }} years in C# and SQL · English, Greek, Russian
Selected work
ABSD — Artifact-Based Software Development
An open-source method and kit I designed and use on my own projects.
Seven agents turn a conversation into requirements, design, architecture, tests and code — each step owned by one agent, checked by another, and proven by git history and tooling rather than good intentions.
ManagerAnalystDesignerArchitectTesterBuilderAuditor
Tests before code
Expected values are derived from the specification and committed as failing tests before any code exists.
Separation of duties, enforced
A write-time hook and a git-history check stop the code author from writing or editing its own tests.
Independent architecture review
A second agent rules on the blueprint before a line is written. It fixes nothing; it only decides.
Evidence rule
The Auditor runs every suite itself at the commit under review — a check that did not run is red.
{{ g.name }}
Let's talk
Email first — it is the one channel I read every day.