Past Work

Systems that came before.

Selected earlier work in game engines, distributed systems, machine learning platforms, reinforcement learning, and applied AI—reconstructed from original artifacts where possible.

2015.09
NVIDIA GTC Korea · Talk + system retrospective

Multi-Agent Game AI Deep Learning Cluster

A production-oriented reinforcement-learning system for cooperative RPG combat: MiniLOL simulation, parallel experience generation, automated competition, hyperparameter search, and continuous deployment.