Roman Kornig

I help product teams ship polished, production-ready web applications end-to-end. With 6+ years of deep React, Next.js, and TypeScript experience, and solid backend skills in Node.js, Kotlin, and Python, I can take full ownership of a feature from database schema and API design through to the final UI.

Where I stand out is the combination of individual depth and leadership experience. At Cybozu, I led a team of 3 to 4 engineers building a production React design system library from scratch: 30+ accessible WCAG-compliant components, a full CI/CD pipeline using Chromatic and GitHub Actions, and a styling migration from Emotion to Vanilla Extract that meaningfully improved runtime performance. That library was adopted across multiple product teams. I been also a core part at building a microservice sales management platform completely end-to-end, covering API design, frontend, customer database, and cloud infrastructure.

I work with AI on two levels. Day-to-day I use Cursor and Claude Code to accelerate delivery without sacrificing architecture or code quality. But I also build with LLMs: my MSc thesis at the University of Amsterdam produced a framework that uses LLM-driven bounded context clustering and a self-healing agentic loop to automate microservice extraction from monolithic codebases. That means I can contribute to AI-augmented products, not just develop faster because of them.

Typical projects I take on: greenfield React or Next.js applications, App Router migrations and frontend refactors, design system and component library work, full-stack feature development with REST or GraphQL APIs, and technical consulting on frontend architecture decisions.

I am based in Amsterdam, fluent in German (native), Japanese (native), and English, and comfortable working with teams across Europe and beyond.