Google Scholar · GitHub · X · Projects · Research

I'm a Privacy Engineer at Google. I'm also interested in learning more about AI privacy & security and understanding model awareness & alignment in general.

Currently trying to answer these questions:

• Does the model really know it is a model?
• What is the understanding of the model about itself?
• Are the model's answers about “escaping” real or inspired from human text?
• What are better questions to think about?

Experience

• Google: helping solve Ads privacy issues
• Enkrypt AI: built LLM guardrails and internal benchmarks
• Skyflow: 2× intern, secure enclave and transient data
• IBM: intern, secrets vault development

Publications

• Privacy‑preserving keystroke analysis using fully homomorphic encryption & differential privacy. ICML TPDP 2021.
• ViT‑inception‑GAN for image colourisation. ICICC 2021.
• Robust adversarial training for detection of adversarial samples. ICICV 2022.
• Measuring the effectiveness of Spinner‑based Randomized‑Response differential privacy communication for sensitive data sharing. SOUPS 2023 (Poster).
• Efficacy of the SAGE‑RT Dataset for Model Safety Alignment: A Comparative Study. NeurIPS Pluralistic Alignment Workshop 2024.
• Robust recovery of adversarial examples. ICML AML 2024.
• VERA: Validation and Enhancement for Retrieval Augmented systems. arXiv 2024.
• SAGE‑RT: Synthetic Alignment data Generation for Safety Evaluation and Red Teaming. COLM 2025 (to appear).

Rand()

• I enjoy playing chess (≈1900 bullet, ≈1700 blitz and have fun playing the Scandinavian Defense sometimes :))
• I'm also interested in philosophy, especially Stoicism.