Fair Math, Inc
- Responsible for all technical decisions in the company — from architecture and technology choices to delivery processes and tooling
- Introduced CI/CD for FHERMA, an FHE challenges platform; migrated the entire company deployment stack to Kubernetes; designed and built verification agents that check user solutions, including auto-scaling; extended verification to support multiple FHE libraries (OpenFHE, Lattigo, Apple Swift Homomorphic Encryption)
- Developed a Cosmos SDK–based blockchain for an FHE computer platform. Key milestone: ran a neural network trained on CIFAR-10 entirely in FHE mode — encrypted image as input, encrypted class label as output, with all inference performed on ciphertext without decryption
- Built an FHE runtime for ONNX models — onnx-fhe-runtime — enabling encrypted evaluation of ONNX computation graphs
- Implemented a proxy that connects Cosmos-based networks to Blockscout explorers — shuttle
- Prototyped private payments for Ethereum based on Pedersen commitments and zero-knowledge circuits written in Circom