Differential privacy, applied cryptography, verifiable computation. 6 peer-reviewed publications.
- dp-diabetes-prediction Differentially private learning on imbalanced clinical data: DP-SGD, an 810-configuration grid, and stability analysis over 100 runs per configuration.
- zkvm-gpu-lab RISC Zero proving and on-chain EVM verification of a satellite least-squares integrity check, with every published number backed by a committed JSON record.
- voting_machine Voting machine prototype in Rust with embedded backdoors, built for red-team analysis of election integrity controls.
- ownft_v2 Self-hostable IPFS-backed data availability for NFTs, with a hardware wallet component. ETHGlobal NFTHACK 2021.
- How SMOTE Quietly Cancels Your Differential Privacy: why oversampling before private training turns a record-level guarantee into a group-level one.
- What It Takes To Build A Differentially Private Mechanism: sensitivity, composition, and the accounting that has to hold end to end.
- Protocol Review of SecureNN: a review of eprint 2018/442, checked against both published revisions and the reference implementation.