I'm Wee Joe, an AI research engineer interested in mechanistic interpretability, reinforcement learning, and emerging paradigms in agentic intelligence, particularly world models, multi-agent coordination, and the challenge of keeping post-AGI systems human-aligned.
I am currently pursuing a BSc in Computer Science at University College London . Earlier, I completed a Diploma in Computer Engineering at Singapore Polytechnic (Apr 2020 to May 2023), where I graduated as valedictorian. I have also taken part in the Stanford ASES Entrepreneurship Bootcamp .
I believe these research directions are among the most consequential of our time. Mechanistic interpretability can make AI systems auditable and trustworthy, a prerequisite for deploying them in medicine, education, and governance. World models and multi-agent coordination unlock AI that genuinely reasons and collaborates, compressing decades of scientific progress into years. And getting alignment right is what determines whether the transition to post-AGI systems expands human agency or erodes it. The stakes make it the most meaningful problem I can work on.