Frontend, done properly.

The thing I do most. React, Next.js, TypeScript. Production-grade frontends for SaaS, fintech, energy, media. Component architecture that won't rot, state management that makes sense, performance that doesn't degrade as the team grows. Ten years of shipping this exact stack at companies like Argenta, Mediahuis, Hexagon and Cognite.

• React / Next.js / TypeScript
• Component libraries & design systems
• State management (Redux, Zustand, RTK)
• Testing (Jest, Playwright, RTL)

End-to-end product work for ambitious SaaS, dashboards, and B2B platforms. From greenfield MVP to scaling an existing product. I've built dashboards for energy traders, banking flows for Argenta, internal tooling for media companies. Same shape of problem, different domain.

• Greenfield product builds
• Dashboards & admin tooling
• Auth, billing, multi-tenancy
• API design & integration

Modern marketing sites and shops, fast and easy to manage. Headless CMS so you can update copy without calling me. Stripe and other payment providers of choice. SEO that doesn't make you wince.

• Marketing sites & landing pages
• Headless CMS (Payload, Contentful)
• E-commerce (Shopify, Medusa)
• Performance & SEO baked in

Already have a team but stuck on architecture, performance, or quality? I review codebases, mentor juniors, run code-review sessions, and help you ship. Drop-in role for a few weeks or ongoing technical advisory, without the long-term commitment.

• Architecture & code reviews
• Performance & bundle audits
• Mentoring & pair programming
• CI/CD & testing strategy

Selective work on Solana. dApp frontends, wallet connectivity (Phantom), integration with existing smart contracts. I take on a small number of these per year alongside my main frontend practice. Honest about scope: I'm a frontend engineer who's worked in the ecosystem, not a smart-contract auditor.

• dApp frontend development
• Solana / web3.js integration
• Phantom wallet connectivity
• UI for existing smart contracts

I use AI tools daily and I'm happy to add LLM features where they make a product genuinely better: streaming chat UIs, RAG over your docs, AI-assisted internal tooling. But it's a complement to a solid codebase, not a replacement. If "we need AI" is the brief, the first call is figuring out whether you actually do.

• Streaming chat & assistant UIs
• RAG integrations
• AI-assisted internal tooling
• Honest scoping