At BRAC University, I’m investigating how CKKS approximate homomorphic encryption affects fairness metrics in privacy-preserving machine-learning audits. I compare plaintext and encrypted calculations and use Python and TenSEAL to evaluate numerical error and possible fairness-threshold changes.
I developed a university advising and registration platform with role-based access control, two-factor authentication, encrypted data storage, and integrity protection. I’ve also implemented filesystem journaling and lottery scheduling in xv6/RISC-V, and developed PRIVHAT, a cryptography CLI for encryption and digital signatures.
