Fixed-term: The funds for this post are available until 31 December 2027 in the first instance.
How can mutually untrusting AI agents, perhaps representing different companies, convince each other that they are trustworthy? How can one convince the other that it is not a fraudster, that it is telling the truth, that it is capable of delivering what it promises, and that it will abide by any contract it enters into?
The thesis of the Centre for Cryptographic Trust Infrastructure (CCTI), recently funded by the Advanced Research + Invention Agency (ARIA), is that cryptographic protocols (zero-knowledge proofs, secure multiparty computation) and trusted hardware (TEEs, tamper-proof sensors) will enable initially untrusting agents to establish trust by proving facts about themselves to each other, without revealing sensitive information. Our approach combines zk-SNARKs with cryptographic attestations about the physical world, and with game-theoretic/economic models of agents' behaviour.
We are looking for an experienced researcher in the area of applied cryptographic arguments of knowledge/zero-knowledge proofs. The post is based at the University of Cambridge Department of Computer Science and Technology. It is a part of CCTI, an international, interdisciplinary collaboration between academic, industry, and nonprofit partners that connects deep technical expertise with exciting real-world use cases.
Potential projects include (but are not limited to):
- Designing, implementing, and verifying the cryptographic protocols for a database that can prove in zero knowledge that its responses to queries are correct. -Working with game theory researchers on cryptographic commitments of agent reasoning, such as proving in zero knowledge that an action was rational under a privately held belief, and that the belief evolved according to a committed update rule,
- Improving mechanisms for establishing the authenticity of data, for example by proving that it was obtained from a server with a particular TLS certificate,
- Adversarial analysis of agent-to-agent protocols: what a sequence of proofs, queries, and refusals reveals about the parties, even when every individual message is zero-knowledge
Role requirements:
- A PhD in Computer Science or a closely related discipline, or an equivalent level of research experience.
- Research experience in designing, analysing, implementing, and verifying the correctness of cryptographic protocols,
- Knowledge of state-of-the-art cryptographic proof systems, such as zk-SNARKs, zkVM internals, arithmetization, lookups, proof recursion and/or folding schemes
If you have any questions regarding the role or the application process, please contact Martin Kleppmann ([email protected]).
Questions about the post and the recruitment process may be addressed to the HR Team at [email protected]
Please quote reference NR50856 on your application and in any correspondence about this vacancy.
The University actively supports equality, diversity and inclusion and encourages applications from all sections of society.
The University has a responsibility to ensure that all employees are eligible to live and work in the UK.
