PETS 2026
Oblivis
Delegated-query protocols and efficient oblivious transfer for private retrieval under a non-collusion assumption, including settings with constrained clients.
Research
Research areas
My research spans cryptographic foundations, protocol design and cryptanalysis, together with implementation and performance evaluation where appropriate. I work on private computation, private data access, verifiable computation, privacy-preserving AI and secure digital exchange.
Precise threat modelsProvable guaranteesMeasured performance
01
Foundations and protocol design
What can cryptographic protocols guarantee, under which assumptions and at what cost?
I develop formal models, constructions and proofs for private set intersection, oblivious transfer, verifiable computation and time-lock puzzles, including post-quantum settings. Across these areas, I study security assumptions, adversarial models and efficiency, alongside implementations where appropriate.
AsiaCCS 2025
Scalable Time-Lock Puzzle
A verifiable delegated framework for time-lock puzzles across clients and servers with unequal computational resources.
02
Cryptanalysis and protocol security
Where do cryptographic constructions fail when their representations, assumptions or proofs hide exploitable structure?
I analyse proposed protocols, identify attacks that contradict or fall outside their claimed guarantees, explain why the original proofs miss them, and develop mitigations where possible. This work includes private set intersection, outsourced computation and distributed protocols.
ESORICS 2021
Polynomial Representation Is Tricky
Attacks on three private-set-intersection protocols previously proved secure against active adversaries, with an analysis of why the proofs missed them and proposed mitigations.
Cryptology ePrint Archive · 2023
Repeated Modular Squaring Service Revisited
An attack that allows servers to receive payment without completing delegated repeated-squaring work, followed by protocol mitigations.
03
Privacy-preserving AI and computation
How can organisations compute across sensitive or distributed data while controlling what is revealed about records, models and intermediate results?
I develop cryptographic methods for privacy-preserving AI, including federated learning, private inference and distributed data analysis. The work considers leakage, malicious participants, constrained devices, data quality and verifiability.
FLTA 2025
STARLIT
Privacy-preserving federated learning for financial fraud detection. STARLIT was a joint first-place winner in the UK-US PETs Prize Challenge.
NDSS 2025
Private data deduplication
Private deduplication across distributed datasets, evaluated for language-model training quality and efficiency.
04
Financial cryptography and secure digital exchange
How can payments and financial collaboration provide explicit privacy, fairness and dispute guarantees?
I work on payment and service-exchange protocols, privacy-preserving financial data collaboration, fraud detection, insured cryptocurrency transactions, and blockchain and smart-contract security.
Financial cryptography
Insured Cryptocurrency Transactions
A framework for insured cryptocurrency transactions and dispute handling in decentralised payments.
IEEE EuroS&P 2023
Recurring Contingent Service Payment
A recurring contingent-payment protocol for delivering verifiable digital services while managing repeated exchange and dispute risks.
For technical readers
Complete publication record
The examples above are a selection. Google Scholar and DBLP provide the complete bibliographic record.