🔐 Multi-Party Computation
From XOR secret splitting to threshold ECDSA wallets. Master the cryptographic protocols that let multiple parties compute together while keeping their inputs private — the technology behind privacy-preserving ML, blind auctions, and cryptocurrency custody.
Why Python?
MPC concepts are language-independent — the math works in any language. Python is recommended because it has the best ecosystem for learning: pycryptodome for primitives, tenseal for FHE, crypten for private ML, and python-ecdsa for threshold signatures. Production MPC systems are typically written in Rust (MP-SPDZ internals, threshold-crypto) or Go (tss-lib). Once you understand the protocols here, porting to any language is straightforward.
🟢 Phase 1: Cryptographic Foundations (1–10)
Secret splitting, modular arithmetic, Shamir sharing, commitments, oblivious transfer, threat models, and the Millionaires' Problem — the ideas every MPC protocol is built on.
🔵 Phase 2: Core Protocols (11–16)
Verifiable secret sharing, replicated sharing, Beaver triples, SPDZ MACs, Yao's garbled circuits, and OT extension — the complete toolkit for 2-party and 3-party MPC.
🟣 Phase 3: Advanced Protocols (17–22)
Paillier homomorphic encryption, LWE and FHE concepts, secure aggregation, private information retrieval, threshold signatures, and OT-based PSI.
🔴 Phase 4: Expert Systems (23–25)
MP-SPDZ framework, CrypTen private neural network inference, and zero-knowledge proofs with MPC-in-the-head. From research tools to production deployments.
⭐ Capstone Projects (26–30)
Five production-grade systems integrating everything in the course: blind auctions, private federated learning, threshold cryptocurrency wallets, genomic privacy, and a full-stack MPC analytics platform.