Patrick Connolly

Algorithms·Privacy·Software Systems

Computer scientist working at the intersection of algorithm research, privacy, and high-performance software systems.

About

I am a final-year Computer Science undergraduate at Trinity College Dublin, currently on year abroad at the University of Tokyo, where I work as a research assistant to Prof. Vorapong Suppakitpaisarn on differentially private fair division algorithms.

Before computer science, I read Law at Cambridge - graduating in 2023 - which continues to shape how I think about systems where correctness, fairness, and accountability matter. My current technical interests sit at the intersection of theoretical algorithms, privacy-preserving computation, and the engineering of high-performance concurrent systems.

Contact

Research

Differentially Private Fair Division of Divisible Items

University of Tokyo · with Prof. Vorapong Suppakitpaisarn

Designing jointly differentially private mechanisms for fair allocation of divisible goods, focused on the privacy-utility tradeoff for envy-free and proportional outcomes. Established improved upper bounds on the algorithmic cost of privacy in indivisible goods setting.

Differential Privacy Fair Division Mechanism Design Convex Optimisation

Education

BA, Computer Science

Trinity College Dublin

First-class results across years one and two; Foundation Scholarship and Book Prize awarded both years. Three-time elected student representative for the School of Computer Science & Statistics.

Visiting Student, Year Abroad

The University of Tokyo

Coursework in mathematical optimisation, algorithms for information security & privacy, programming languages theory, erorr and noise handling, and numerical simulation.

BA, Law

University of Cambridge

Upper-second-class honours. Concentrations in international and commercial law.

Selected Work

Limit Order Book & Matching Engine

C++ · Concurrency · Market Microstructure

High-performance matching engine implementing price-time priority, partial fills, and the standard order types found in equities and crypto venues. Targeting cache-friendly data structures and sub-microsecond per-operation latency on a single thread.

Concurrent Reservation System

Python + FastAPI · Postgress · Production Deployment

High-throughput booking engine designed for a real sports club deployment, with a focus on correctness under contention - atomic compare-and-swap for slot acquisition, a write-ahead log for crash recovery, and integration testing under simulated concurrent load.

SynthetiQ - Privacy-Preserving Data Pipeline

TCD Software Engineering Final · Mentored by Millennium Management

Led an eight-developer Scrum team to a competition final from a starting field of forty. The system handles streaming ingest, GDPR-compliant redaction, and synthetic data generation, with end-to-end authentication and access control.

Bayesian State Estimation for a Hydrofoil Solar Vessel

University of Tokyo Solar Boat Project

Implemented an Unscented Kalman Filter for non-linear Bayesian state estimation of GNSS data, with a constant-turn-rate motion model, innovation-based certainty gating, and WGS-84 → ECEF → ENU coordinate transforms. Contributed to a second-place finish at the Japan National Solar Boat Championship.

ARM Assembly Optimisation

Trinity College Dublin · Computer Architecture Coursework

Achieved top memory- and cycle-efficiency results across the cohort on prime detection, bit-pattern recognition, and constrained search problems through aggressive branch pruning and instruction-level optimisation.