kateryna

Conducted research on privacy-preserving analysis of encrypted visual data using homomorphic encryption techniques. Explored the application of AI models to encrypted data, focusing on secure inference using homomorphic encryption schemes. Investigated approaches for combining machine learning with cryptographic techniques to enable secure processing of sensitive data.

Developed and optimized ZK circuits using Circom and Plonky2 to enable verifiable computation within a decentralized differential dataflow framework. Implemented performance-critical cryptographic primitives and state-tracking components in Rust to accelerate proof generation. Analyzed and integrated SNARK/STARK constructions to support secure, scalable, and privacy-preserving data processing.

Designed and implemented a high-performance ZK state transition verifier for NEAR Protocol using the Plonky2 framework, tailored for the Wormhole light client ecosystem. Developed core cryptographic circuits in Rust to verify proof-of-stake consensus and header updates for cross-chain interoperability. Conducted R&D on optimizing proof generation and verification constraints, bridging cryptographic primitives between NEAR and external networks.

Contributed to the implementation of a digital signature scheme based on the post-quantum FALCON algorithm. Implemented algorithms related to number-theoretic transforms (NTT), FFT-based polynomial arithmetic, and NTRU structures. Worked on cryptographic primitives used in lattice-based post-quantum cryptography. Participated in testing and evaluation of cryptographic implementations for correctness and performance.