Full-stack Software Engineer — Game engines, real-time backends, native apps, AI in production
I build complete systems rather than slices of them - an MMO client in Unreal, a voice agent running on the customer’s own hardware, a tournament platform that takes its scores out of the running game.
Open to new work — Vienna, Austria
Selected work
2025 — Deployed at customers
AI telephony that runs entirely on the customer’s own hardware
A voice agent that answers the phone for doctors’ practices, law firms and tax advisors, running entirely on hardware in their own building.
Frostphone, in detail — one call, end to end
The number the customer already publishes, over a SIP trunk into their own PBX. Nobody has to be told to ring a different line.
The call becomes a real-time media stream. Everything after this point is on a clock, because a stall is not a slow page - it is somebody holding a phone and hearing nothing.
This is the step where a cloud product would post the audio to somebody else’s API. There is no such call to make here: the model runs on a GPU in the customer’s own building.
It gets whatever VRAM the two speech services leave behind - six gigabytes of a twelve-gigabyte card - and a bigger card simply buys a stronger model. Not a reasoning model, though: thinking tokens reach the caller as silence.
Neural synthesis, back down the line the call came in on. The caller hears an answer, and the audio never left the building.
Twelve containers, all of them on hardware the customer owns. On a phone line the latency budget is the specification, not a target.
The rest of it
2019 — Shipped
Four years owning every raid, dungeon and boss encounter on the largest Classic+ server in the world
C++ / SQL / VMaNGOS / MySQL / Encounter scripting
2026 — In development
An MMO client written from scratch in Unreal Engine 5
C++ / Unreal Engine 5 / UMG / Python / Binary protocol
2021 — Working
A reinforcement learning policy driving a live Rocket League client at frame rate
Python / PyTorch / TorchScript / C++ / Binary IPC
2025 — In production
Self-hosted team communication with end-to-end encryption
TypeScript / React / Node.js / tRPC / MediaSoup
2026 — In development
Tournament platform that reads the live score out of the game itself
TypeScript / Fastify / Next.js / Socket.IO / Prisma
2026 — In development
An Android music player with no backend of its own
Rust / Tauri 2 / Vue / TypeScript / Kotlin
Stack
I build and operate systems around models: local inference, real-time pipelines, latency and VRAM budgets, and what happens when the model is wrong. I have not trained foundation models.
Hiring
Vienna, working with teams across Europe and the US. I am useful anywhere between a rendering loop and a deployment pipeline. Email is the fastest route and I answer within a day.