Research
My research makes machine learning and security analytics work on encrypted data, so organizations can detect threats and collaborate on models without exposing sensitive information. Core themes: privacy-preserving machine learning, fully homomorphic encryption (FHE), and security analytics.
Secure and trustworthy data science
My work innovates at the nexus of applied cryptography and privacy-preserving machine learning, resolving the tension between the power of data analytics and the imperative of individual privacy, with deployable solutions for healthcare, finance, and critical infrastructure.
Cryptographic primitives for computation
Secure Multi-Party Computation (MPC) and Fully Homomorphic Encryption (FHE) optimized for machine learning, with zero-knowledge proofs for provable integrity on encrypted data.
Provable privacy in decentralized systems
Communication-efficient, differentially private Federated Learning, combining Secure Aggregation with Differential Privacy to resist inference attacks.
Ethical & applied PPML
Translating cryptographic guarantees into clinical data sharing and secure financial modeling, with governance for fair, responsible deployment.
Current and recent projects include Fairis (improving group fairness in private collaborative ML), privacy-preserving tax analysis, and local-government supply-chain cybersecurity.
Selected publications by topic
Below is a selection of peer-reviewed publications, with author name shown in burgundy. The counts above link to the complete, continuously updated record across profiles.