Jatan Loya

I'm interested in researching and developing privacy and security solutions for billions of users amongst other interests.

I'm currently learning more about AI, particularly privacy and security of these systems. Trying to answer questions like:

1. Does the model really know it is a model?
2. Does the model know we are "testing" it?
3. Are the model's answers about "escaping" real or inspired from human text?
4. What are better questions to think about?

I am currently a Privacy Engineer at Google, solving privacy issues at scale.
I have worked on building AI privacy and security solutions at Enkrypt AI. I also completed two internships at Skyflow and IBM.

My previous research includes:

• Keystroke biometrics
• Privacy-preserving machine learning
• AI Guardrails
• Fully Homomorphic Encryption

I like playing chess. Currently ~1900 bullet and ~1700 blitz. Also I like the Scandinavian!

I am interested in reading philosophy, particularly Stoicism amongst other fun stuff.