Will Gilroy Email & Phone Number

Seattle, Wa, Us

Build, Design, and Maintain REST APIs (OData) for OneDrive and Sharepoint to hundreds of millions of customers.Bespoke API collaboration to enable flagship products like Microsoft Project Nucleus and Microsoft Graph in OneDrive and Sharepoint.

Researching Numerical Methods and Probabalistic Algorithms for Large SystemsSkills in: Data Science, Optimization, Statistical Theory, Programming in PythonNovel Algorithmic Approches to Ranking Sports Data. Invention of Weighted Extended Randomized Kaczmarz Algorithm; improvements in error detection and supression. Computational experiments for real and synthetic data

Researched applications of Graphs Neural Networks to natural langugage problems in crime fiction. IEEE Big Data Publication for novel Knowledge Graph analysis technique.Skills: Neural Networks, Graph Theory, Natural Language Processing, programming in PythonI worked as a Mathematics Researcher building Knowledge Graphs for solving natural language crime fiction. I had some experience building natural language systems and implemented graph neural network procedures. We had a publication in IEEE Big Data, 2020 Publication (https://arxiv.org/abs/2011.11804) for developing a new knowledge graph analysis technique which improved costs over other techniques.

Greater New York City Area

Project: Network Orchestration Development Environment (Technology Infrastructure)Skills: Full-Stack web-development, Dev-Ops, Network Orchestration, Computer Networks.I worked as a Software Engineer on the Network Orchestration team at Bloomberg. I spent this summer developing a full-stack IDE application for building Network Orchestration procedures. Within this role I learned software engineering team procedures. My project significantly increased developer efficiency and program robustness.Technologies: Python: Flask, SQLAlchemy; Javascript: React, Redux, Sagas; Linux Red Hat.Technology Infrastructure, Network Orchestration.

Neural Networks and Harmonic Analysis.Do Neural Networks 'learn' to do Harmonic Analysis? Rapid prototyping with Pytorch, Jupyter Notebooks, and Scientific Computing Libraries. We were interested in understanding the internal structure of Neural Networks once they are trained. As a lens into this investigation, we trained a Neural Network to classify Circular Distributions. Once trained, we compared the internal Fourier Series structure to the results given by Harmonic Analysis. Harmonic Analysis tells us that you can classify circular distributions by focusing your attention on particular Fourier coefficients -- we wanted to know if our neural network would 'do the same' when classifying Circular Distributions, or if it would 'do' something else entirely.

Selected over 1000+ applications to be member of the Cohort 6 of the Sutton Trust US programme. The Sutton Trust supported me to apply to universities in US and is run by the US-UK Fulbright Commission and the Sutton Trust.A fantastic experience; to them I owe much of the joy of my undergraduate years.