CV
Academic CV. Download the PDF using the button above, or edit _data/cv.yml to update the content.
Contact Information
| Name | Eddie Cunningham |
| Professional Title | PhD Candidate |
Professional Summary
PhD candidate at UMass Amherst researching the intersection of differential geometry, generative modeling, and representation learning.
Experience
-
Sep 2019 - Present Amherst, MA, USA
Graduate Research Assistant
University of Massachusetts, Information Fusion Lab
- Published papers at top conferences on normalizing flows and dimensionality reduction
- Implemented all research from scratch in JAX, building three open-source libraries for diffusion/flow based generative models, stochastic processes, and Riemannian geometry
- Applied methods across problem domains including self-supervised representation learning, time series generation, and density estimation on manifolds
-
Jun 2021 - Aug 2021 Menlo Park, CA, USA
Research Intern
SRI International, Computer Science Laboratory
- Led research project on finding principal structure in normalizing flow models
- Developed theory and algorithms to extract low-dimensional structure from normalizing flow models
- Intern project resulted in paper accepted at ICML 2022
-
Jul 2017 - May 2019 Olathe, KS, USA
Sensors and Algorithms Team / GPS Positioning
Garmin
- Used sensor measurements from wrist mounted watch to develop Bayesian machine learning models
- Developed real-time routing and positioning algorithm to build optimal race pace segments
- Built Python testing framework for C++ algorithms, drastically reducing algorithm development time
-
May 2016 - Aug 2016 Cupertino, CA, USA
Software Engineering Intern
Apple, Apple Watch Team
- Developed automated testing framework for Apple Watch software, reducing manual QA effort
- Implemented features in Objective-C, Swift, and Python for the Apple Watch software platform
Education
-
2019 - 2026 Amherst, MA, USA
M.S./Ph.D.
University of Massachusetts Amherst
Computer Science
- Thesis: Orthogonal coordinates for representation learning
- Advisors: Daniel Sheldon and Madalina Fiterau
-
2013 - 2016 Ann Arbor, MI, USA
Software
Research Areas
Publications
-
2026 Conformal Coordinate Frames for Disentanglement
Under review, ICLR 2026 Workshop on Geometry-grounded Representation Learning and Generative Modeling (GRaM)
Proposed conformal coordinate frames as a principled geometric approach to disentangled representation learning.
-
2025 On Autoregressive Time Series Generation using Flow-based Generative Models
Preprint
Developed flow-based generative models for autoregressive time series generation.
-
2022 Principal Component Flows
International Conference on Machine Learning
Introduced principal manifold flows for density estimation on manifolds with variable dimensionality.
-
2021 A Change of Variables Method For Rectangular Matrix Vector Products
International Conference on Artificial Intelligence and Statistics
Developed a change of variables method for rectangular matrix vector products in flow based models.
-
2020 Normalizing Flows Across Dimensions
ICML Workshop on Invertible Neural Networks and Normalizing Flows
Proposed methods for extending normalizing flows across dimensions.
-
2018 Explainable Genetic Inheritance Pattern Prediction
NeurIPS Workshop on Machine Learning for Health
Presented explainable methods for inheritance pattern prediction with structured probabilistic models.
Awards
-
2019 Spaulding-Smith Fellowship
University of Massachusetts Amherst