The person building Grundval
Ten years of full-stack engineering, now pointed entirely at one dark fantasy deckbuilder.
I've shipped production software in both small teams and large organizations, usually owning the parts where a mistake is expensive: payments, data correctness, systems under real load.
Recent work has centered on AI-powered applications, system architecture, performance, and developer experience.
I take on problems where the right answer isn't obvious yet, and I care whether the result actually holds up in production.
Senior full-stack engineer with ten years of experience, mostly in FinTech and SaaS. I've spent my career building products within other companies. Now I'm putting that experience into building something of my own.
It started on paper. Before a line of code existed, the combat was index cards on a table — draw a hand, commit a combo, remove cards as damage, see whether the loop was any good. It wasn't, for a while. Deck-as-health only became interesting once losing cards started costing you the specific thing you'd been saving them for, and that took a dozen sessions of shuffling paper to find. Everything since has been implementation.
A deckbuilder is an unreasonable number of disciplines for one person. Engine, combat balance, art direction, worldbuilding, marketing copy, a website, three languages of localization. The traditional answer is a studio. The answer I'm testing is a pipeline: written specifications first, generation as a drafting step, and a person deciding what actually ships.
That pipeline is the real experiment. React and Next.js render the game, NestJS serves the API — but the engine underneath is plain TypeScript with no framework dependency at all, so the rules are testable in isolation and the bugs are reproducible. The lore lives in version control with a governing document that defines what the game is permitted to assert — the short version is nothing, because every fact in this world reaches the player through someone with a motive. The heraldry has a proper blazon, because if you're going to invent a kingdom's coat of arms you may as well get the grammar right. Character art went through four rounds of direction before anything was usable, which is roughly the ratio I'd expect from a human illustrator.
None of that is the interesting part on its own. The interesting part is that a decade of engineering habits — specs, version control, reproducible builds, honest scoping — turns out to transfer almost completely to creative work, and the tooling has finally caught up enough to make a solo attempt reasonable rather than delusional.
Hollowfort is the first chapter. Grundval is the series. The public site and the playable beta are where the experiment meets players.
For playtest notes or professional questions, LinkedIn and email both reach me. Feedback from a finished Hollowfort run is especially useful.