AI in production
At RUTILEA I lead a system migration for Japanese government agencies and manage a 128-node H100/H200 cluster running self-hosted LLMs.
Privacy and AI
I'm Yumi, a forward-deployed engineer with a cybersecurity background.
I build AI systems for institutions, research LLM safeguards, and care about keeping people in control of their data.
selected work
LLM safeguards
As a CyberVSR research assistant at Rochester Institute of Technology, I studied jailbreaks and the protections models use against harmful prompts.
Why privacy matters
A TEDx talk on digital privacy.
writing
How I think about privacy.
My personal thoughts on privacy.
I Asked My AI to Get Me a Restaurant Reservation
There's a yakitori spot in Kyoto that's nearly impossible to book. Their slots open unpredictably and vanish fast. So I built a bot to watch for me.