Craig Gentry

Fully homomorphic encryption (FHE) allows arbitrary computations to be performed on encrypted data without ever decrypting it. Craig Gentry constructed the first fully homomorphic encryption scheme in 2009 while a PhD student at Stanford, solving a decades-old open problem that had been posed by Rivest, Adleman, and Dertouzos in 1978.

Gentry's scheme uses lattice-based cryptography and relies on a technique called 'bootstrapping,' which refreshes the ciphertext after each operation to prevent noise from accumulating and corrupting the result. Bootstrapping involves homomorphically evaluating the decryption circuit on the encrypted ciphertext itself.

Early FHE schemes were extremely slow, but subsequent improvements by Gentry and others have reduced the overhead dramatically. Modern libraries like Microsoft SEAL, IBM HELib, and Google's FHE C++ library make certain FHE operations feasible for privacy-preserving analytics, though they still carry significant computational costs compared to unencrypted computation.

Gentry received the ACM Doctoral Dissertation Award in 2009 and was recognized by MIT Technology Review's TR35 list of young innovators. His FHE work is considered one of the most significant breakthroughs in theoretical cryptography in the 21st century.