MY EVERYDAY COMPANIONS

Life, with "a little" help.

I’m Bo. Meet the little crew I think, remember, and build with.

THE HUMAN IN THE LOOP

Hi, I’m Bo.

I’m a PhD student at Indiana University, advised by Thai Le. I explore AI while using it in everyday life. This is my little circle of tools for catching ideas, thinking things through, and turning them into something real.

I bring the questions. My companions help me take the next step.

RESEARCH / WORKING QUESTIONS

Taste Is All You Need.

I study how AI systems behave when the benchmark ends and the world gets messy.

Unlearning that holds up

Machine unlearning, knowledge protection, ASR unlearning, and the gap between genuine forgetting and superficial behavior suppression.

LLMscausal evidencerobustness

Agents in the physical world

Embodied-agent evaluation, grounding, action validity, long-horizon planning, and the unlearning gap between language models and physical systems.

embodied agentsfailure analysisphysical AI safety

Evidence before confidence

Agent evaluation harnesses, mechanistic interpretability, and audio-language-vision systems that make model behavior easier to inspect.

evalsinterpretabilitymultimodal systems

TASTE IS ALL YOU NEED

Good research is not just a bigger model or a longer benchmark. It is choosing the right question, finding the failure mode, and building evidence that survives contact with reality.