From Novice to Starlight King: How AI-Driven Game Design Rewrote My Destiny in Los Angeles

I didn’t grow up chasing idols—I grew up debugging them.

At 22, I walked through Shibuya’s glowing stalls like a debug session: every bet was a state transition in a Markov chain. The crowd screamed for ‘lucky numbers’—but I saw probabilities. My CS degree taught me that ‘winning’ wasn’t an event—it was an optimization problem.

I built my first game using Unity and reinforcement learning models. Each spin? A reward function with entropy. Each ‘flip’? A gradient descent toward expected value. No mystical rituals—just clean code, calibrated stakes, and disciplined time budgets (20–30 minutes max).

The ‘Starlight Casino’ isn’t about luck—it’s about latent space exploration. Players think they’re dancing under neon gods—but I’m training agents to predict when the next bonus triggers. My third title? A procedural generation of player behavior patterns across cultures.

I stopped chasing jackpots. Started tracking win rates instead.

In LA, we don’t gamble—we simulate decision trees under ambient light.

Your next move? Don’t roll the dice. Click the button. Let the algorithm decide.