See the system, not just the symptom.
Start with real behavior, constraints, incentives, and feedback loops. The bug is often only the most visible part of the problem.
Staff EngineerFormer Engineering Manager
I lead AI-enabled development and design the quality, delivery, release, and incident systems that help teams move with confidence.
How I think
A working belief
My work starts where code meets reality: how decisions are made, how feedback arrives, how risk becomes visible, and what happens when production disagrees with the plan.
Start with real behavior, constraints, incentives, and feedback loops. The bug is often only the most visible part of the problem.
Evidence creates a shared language. When tradeoffs are visible, teams can make deliberate decisions instead of optimistic guesses.
Hotfix, rollback, observability, and incident response are not afterthoughts. They are part of how trustworthy software is designed.
AI should accelerate review and execution while keeping ownership explicit. Fast output only matters when judgment stays intact.
Systems I build
Tools matter. The operating model around them decides whether they create leverage or noise.
Agent-based development tooling, automated code review, and workflows designed to work reliably for both people and machines.
Agents · review · workflow designCI/CD, unified deployments, release processes, and practical hotfix and rollback paths that make change safer to operate.
GitHub Actions · releases · recoveryEnterprise testing platforms, team-embedded quality assistance, flaky-test management, coverage, and maintainable end-to-end systems.
Automation · evidence · reliabilityDistributed team leadership, hiring, coaching, incident response, and operating models that connect Engineering, Product, and Design.
Teams · systems · shared ownershipSelected experience
More than a decade across automation, quality leadership, management, delivery, and staff-level engineering.
2026 — now
Leading an Engineering, Product, and Design transition to AI-assisted development through automated review, agent workflows, and the codebase and pipeline structure that supports them.
2022 — 2026
Built and led a five-person distributed team. Reworked delivery around GitHub Actions, unified deployments, release and incident response, and reliable hotfix and rollback paths.
2021 — 2022
Introduced team-embedded quality assistance, flaky-test management, a code coverage reporting mechanism, and more stable, maintainable end-to-end testing systems.
2015 — 2021
Built test frameworks, hired and trained quality engineers, introduced self-contained CI/CD solutions, and developed a lasting focus on transparency, adaptability, and reflection.
Professionalism
Professionalism is not polish for its own sake. It is making constraints visible, questions safe, ownership explicit, and recovery routine—then doing the work well enough that others can rely on it.
Hungary · working globally