Suraksh AI

Award-Winning Research

Our founders have accrued more than 7000 academic citations combined through their peer-reviewed research work. Here are some canonical examples:

PrivaCT: Circuit Privacy in Constant Time for FHE Evaluation of Functions

Under Review

Novel approach to achieving circuit privacy in constant time for fully homomorphic encryption evaluation.

SecureRAG: End-to-End Secure Retrieval-Augmented Generation

Under Review

Secure retrieval-augmented generation system that maintains privacy throughout the entire pipeline.

CryptoFace: End-to-End Encrypted Face Recognition

CVPR 2025

The first end-to-end encrypted face recognition system with up to 8x speedup via FHE-aware network design.

Practical Biometric Search under Encryption: Meeting the NIST Runtime Requirement without Loss of Accuracy

IEEE Transactions on Biometrics, Behavior, and Identity Science 2025

A practical approach to biometric search under encryption that meets NIST runtime requirements without sacrificing accuracy.

AutoFHE: Automated Adaption of CNNs for Efficient Evaluation over FHE

USENIX Security 2024

An automated compiler for efficient CNNs over FHE, outperforming Zama and CryptoLab baselines.

Improved Multiplication-Free Biometric Recognition under Encryption

IEEE Transactions on Biometrics, Behavior, and Identity Science 2024

A novel approach to biometric recognition under encryption that eliminates the need for multiplication operations.

BEST PAPER AWARD

HERS: Homomorphically Encrypted Representation Search

TBIOM 2022

Breakthrough homomorphically encrypted biometric feature search algorithm scaled to 100+ million gallery size.

BEST PAPER AWARD

HEFT: Homomorphically Encrypted Fusion of Biometric Templates

IJCB 2022

The world's only homomorphically encrypted biometric template fusion algorithm.

Secure Face Matching using Fully Homomorphic Encryption

BTAS 2018

A first approach to secure face matching using fully homomorphic encryption.

Privacy-Preserving Visual Learning Using Doubly Permuted Homomorphic Encryption

ICCV 2017

A novel approach to privacy-preserving visual learning using doubly permuted homomorphic encryption.