Description
Secure Multi-Party Computation (MPC) has evolved to the point where advanced protocols enable practical, low-latency applications. However, many communication and multimedia tasks—such as message processing, collaborative filtering, private content analysis, and encrypted signal processing—still require custom MPC protocols and strong performance optimization.
This internship focuses on designing and implementing practical MPC protocols for modern communication and multimedia scenarios. The emphasis is on efficient, specialized building blocks that reduce communication costs, increase throughput, and enable near-real-time performance. Potential directions include: privacy-preserving processing of communication workloads, encrypted audio or media signal processing, efficient MPC for AI inference or filtering.
The goal is to take MPC from "theoretically feasible" to practical use through protocol innovation, system tuning, and implementation in existing MPC frameworks such as SPDZ, ABY, MP-SPDZ, or proprietary variants, including evaluation under realistic network conditions.
Duration: flexible, to be agreed (typically 3-5 months), starting time flexible
Responsibilities
* Explore MPC protocols suitable for communication, AI, and multimedia workloads, and identify those that best match the project needs.
* Design and implement protocols or algorithmic components to meet specific application requirements.
* Build a prototype in an MPC framework and benchmark its performance under realistic conditions.
* Analyze results, compare protocol choices, and identify optimization opportunities.
* Summarize the findings which ideally lead to a research publication.
Qualifications
* We are looking for a candidate who is enthusiastic about protocol design and secure computing:
* PhD candidate in computer science/engineering.
* Strong interest in MPC, privacy-preserving computation, or distributed cryptographic protocols.
* Ability to design new protocols or adapt existing MPC constructs to specific workloads.
* Solid implementation skills (e.g., Rust, C/C++, Go, JavaScript, or Python).
* Performance-oriented mindset (communication/computational tradeoffs, batch processing, vectorization).