encrypted-inference

Research prototype for real Mamba-2-130M inference under CKKS/FHE (OpenFHE/FIDESlib-GPU): +0.12% PPL surrogate and a verified 24-layer, 3-token encrypted B300 path.

The open toolkit for LightChain AI: TypeScript SDK for encrypted on-chain inference, self-custodial wallet with built-in AI chat, one-flow worker onboarding, and live network + DAO dashboards.

Product and GTM docs for MPC-secured agent payments and encrypted inference.

Decision-integrity-aware direct synthesis and validation of CKKS configurations.

Runnable LightChain AI inference examples (lightnode-sdk)

Privacy-preserving machine learning with Fully Homomorphic Encryption (FHE): theory plus encrypted-inference demos using Concrete-ML (Iris, sentiment, image filtering). Cryptography coursework (M.Sc. IT, HUTECH).

Browser-based CKKS approximate FHE demo — encrypted floating-point arithmetic, homomorphic neural network inference, rescaling, and the complete FHE trilogy (TFHE + BGV/BFV + CKKS). Part of the crypto-lab suite.