Wesley Maddox
Quantitative Researcher
New York, New York, United States
Summary
🤩
Rockstar🎓
Top SchoolWesley Maddox is a quantitative researcher in New York with a decade of experience applying probabilistic machine learning, deep learning, and Bayesian methods to real-world problems. Currently at Jump Trading, he brings a strong research-to-production background shaped by internships at Facebook and Amazon and a PhD in Data Science from NYU. An active open-source contributor, Wesley has made substantive back-end contributions to flagship PyTorch libraries—GPyTorch and BoTorch—improving Gaussian process stability, memory use, and multi-objective Bayesian optimization. His training spans rigorous statistics and systems biology, blending theoretical depth (MS in Statistics, visiting mathematics at Oxford) with practical engineering that tightens numerical correctness and test coverage in complex ML codebases.
11 years of coding experience
1 year of employment as a software developer
Visiting Students Programme, Mathematics, Visiting Students Programme, Mathematics at University of OxfordBachelor of Science - BS, Systems Biology, Bachelor of Science - BS, Systems Biology at Case Western Reserve UniversityDoctor of Philosophy - PhD, Data Science, Doctor of Philosophy - PhD, Data Science at New York UniversityStatistics, Statistics at Cornell University
English
Github Skills (24)
unit-testing10
pytorch10
python10
model-driven10
machine-learning10
gaussian-processes10
model-building10
bayesian10
optimisation10
gpytorch10
linear-algebra10
modeling10
model-driven-development10
optimization10
decomposition9
Programming languages (7)
Github contributions (5)
cornellius-gp/gpytorch
Sep 2019 - Jun 2022
A highly efficient implementation of Gaussian Processes in PyTorch
Role in this project:
Back-end Developer
Back-end DeveloperContributions:104 reviews, 194 commits, 58 PRs in 2 years 8 months
Contributions summary:Wesley's commits primarily involve modifying and enhancing the core functionalities within the GPyTorch library. The commits demonstrate the addition of new methods, such as those related to Kronecker product and distribution sampling. The user also implemented features to enhance the flexibility and stability of the library, specifically concerning the handling of data types and the management of the log determinant calculation, which are crucial to the project's overall functionality. In addition, the user made various updates to the unit tests, improving the library's quality control by ensuring the accuracy of the mathematical computations.
gaussian-processespytorchgpu-acceleration
meta-pytorch/botorch
Nov 2020 - Apr 2022
Bayesian optimization in PyTorch
Role in this project:ML Engineer
ML EngineerContributions:21 reviews, 18 commits, 23 PRs in 1 year 5 months
Contributions summary:Wesley made multiple contributions focused on improving the functionality and performance of the Bayesian optimization library. They added a new option to the `qMVES` acquisition function to support a wider range of GPyTorch models and enabled gpytorch settings to override botorch defaults. The user also addressed memory inefficiencies in the `HigherOrderGP` model, and added a new multi-objective test problem to expand testing capabilities. Furthermore, they integrated support for input and outcome transforms within the MTGP models.
bayesianpytorch
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.