I spent four years running Infima Games, my own game-dev tooling company, building systems that let non-technical people do technically complex things. The role you're describing (empowering non-technical users to create complex interactive content through AI, removing the traditional barriers) is the same problem with AI as the new primitive instead of game-dev SDKs.

For the past year I've been shipping production AI systems. At TwinMind, an AI memory startup, I built backend systems for real-time voice capture and retrieval at scale. In parallel, I've been building AI Nexus, my own agent gateway for orchestrating persistent agents with multi-layered memory. Context and retrieval quality really matter when the agent runs continuously, because every reply builds on the last one.

Claude Code and Cursor are my primary tools, and I treat output validation and eval design as core engineering work, not an afterthought. I've built eval suites specifically to benchmark memory retrieval accuracy across system changes, because shipping AI features without measurement is guessing. I also run multiple agents on my own infrastructure, so I hit the real production failure modes myself: context drift, retrieval degradation, latency, cost.

The combination you're describing (small team, AI-native workflow, high ownership, fast iteration) is where I do my best work. I'm based in Spain on CET, cleanly within Europe for remote collaboration.

Resume: https://notion.octaviantocan.com/Octavian-Tocan-Resume-34a3c065308b8188b461f46f59886a28

Book a Meeting: https://calendar.notion.so/meet/octaviantocan/lfvw4pfz