ivan@host:/home/ivan$ whoami
Ivan Viragine
ivan@host:/home/ivan$ cat about.txt
AI Engineer with 18+ years of experience building production-grade software systems,
and in recent years, AI platforms and multi-agent architectures in regulated environments.
Focused on evaluation-first engineering, observability, and long-term ownership past launch.
I design systems that survive contact with users: measurable behavior,
debuggable workflows, and platform primitives that other teams can build on.
ivan@host:/home/ivan$ cat scope.txt
- Hands-on Staff/Architect IC scope: platform architecture, reliability, evaluation
- Multi-agent voice and text systems in production
- RAG-backed experiences with automated quality metrics and test harnesses
ivan@host:/home/ivan$ cat AGENTS.md
AGENT BEHAVIOR AND OPERATING CONSTRAINTS
- Evaluation is a first-class capability, not a post-hoc metric.
- Non-determinism must be observable, bounded, and debuggable.
- Agents should fail explicitly and early rather than degrade silently.
- Production behavior matters more than benchmark performance.
- Every agent interaction should be traceable and reproducible.
- Shipping is the beginning: agents must evolve safely post-launch.
- Prefer boring interfaces and explicit contracts over clever internals.
- Usability is part of correctness, not an afterthought.
ivan@host:/home/ivan$ ls links/
github linkedin email
ivan@host:/home/ivan$ cat links/github
https://github.com/ivanviragine
ivan@host:/home/ivan$ cat links/linkedin
https://www.linkedin.com/in/ivanviragine/
ivan@host:/home/ivan$ cat links/email
ivan (at) ivanviragine (dot) dev
--------------------------------------------------------------------------------
Static page by design · ❤️