About the Role
The client runs a mobile feed made entirely of AI-generated browser games created from prompts. Its generation pipeline is the product.
Responsibilities
•Own the full pipeline, from intent extraction and model routing to generation, validation, repair, and runtime
•Improve the playability pass rate and reduce cost per generation at scale
•Run benchmark suites and LLM-as-judge evaluations
•Reduce inference costs through routing, caching, and distillation
Requirements
•Have shipped text-to-game or prompt-to-running-app technology at scale
•Have owned the generation, validation, repair, and evaluation loop, with measurable pass-rate and cost results
•Be a strong engineer with runtime experience in Three.js, Phaser, Babylon, or procedural generation
Conditions
•London-based; remote until October 2026, then relocation
•Compensation is open and top of market for the right candidate
