Opening up the inner workings of AI systems.

Interpretability tooling, sparse autoencoders, activation patching, and AI safety governance.

A solo research practice in Singapore, working worldwide, with a studio side building intelligent installations.