Zero Knowledge Proofs (ZKPs) have become a fundamental building block for creating scalable and privacy-preserving decentralized applications. ZKPs enable verifiable computing, but the powerful math behind ZKPs is both compute and memory intensive. Furthermore, no single proof system excels at efficiently proving the diverse set of operations involved in a ML-based data processing pipeline.

In this project, you will explore producing efficient ZK proofs for the entire data processing pipeline. The project will involve exploring the latest ZK proof systems & lookup techniques and applying them to common pre-processing tasks and ML inferencing tasks used in text, image, and audio processing pipelines. The focus will be on identifying which proofs systems are best suited for the data processing operations involved. Furthermore, you will also explore ways of composing proofs from different proving systems in a universal and extensible manner.
 

Student enrolled in  Ph.D. Computer Science/Engineering. Good system building skills. Interest in, or experience with, ZKP techniques and frameworks Affinity with Machine Learning Language skills: English You will be expected to get up to speed with a variety of ZKP proof systems and frameworks You will bring together multiple open-source projects into a working ZKP system for machine learning and data processing at the edge. You will implement a working prototype and be involved in writing an academic paper related to the project.

Duration: flexible, to be agreed (typically 3-4 months), starting time is flexible

Location: Stuttgart (Germany)