Julien Millet
AI engineer by training.
Game industry executive by practice.
I'm Julien Millet. AI engineer by training, game industry executive by practice, and increasingly convinced that these two worlds have more to say to each other than either suspects.
I trained as an AI and cognitive science engineer in Paris at the turn of the millennium, before deep learning existed as a field, when artificial intelligence still meant symbolic systems, reasoning under uncertainty, and the hard question of how minds model the world. That foundation shaped how I think about the limits of purely statistical approaches, and why the composite AI challenge — making rule-based and neural systems genuinely cooperate — feels to me like the most important open problem in the field right now.
I then spent 20 years building games, founding studios, shipping 23 titles across console, PC, VR and mobile, scaling teams from 5 to 60 people, navigating the full arc from creative vision to production reality. One of those companies made the Deloitte Fast 50. One shipped Solasta, which won France's Pégase award. I also co-founded Game Only, a regional cluster of 100+ game organizations, where I worked on industry financing and ecosystem development at a structural level.
Games are not entertainment software. They are interactive simulations.
The problems our industry has spent decades solving — real-time state management, agent behavior under uncertainty, controllable interaction loops, the coexistence of rule-based systems and learned behavior — are foundational to what the next generation of AI needs to tackle in robotics, industrial automation, and healthcare. That bridge is underexplored, and it is the reason this site exists.
Edge Obs is a field journal. I write about AI architecture, composite intelligence, the hard problems at the intersection of simulation and machine learning, and what it takes to move research from theory into systems that actually work under operational constraints. The posts here are not summaries of papers I read. They are observations from someone who has been building at this intersection for a long time, and who still has more questions than answers.
If you are building at the frontier of AI, if you work in games and wonder what comes next, or if you are trying to understand what intelligent systems actually require to function in the real world — this is written for you.