My work sits at the intersection of cosmology, statistics, and software. I build methods and tools that make it easier to extract reliable science from astronomical surveys and imaging data.

Current: Rubin/LSST supernova cosmology

Probability of selection given cosmology heatmap vs w0 and Omega_m. Example corner plot showing a fit to mock supernova data.

As a Canadian Rubin Fellow at the University of Toronto, I am developing a fully Hierarchical Bayesian Inference framework for supernova cosmology with the Vera C. Rubin Observatory’s Legacy Survey of Space and Time (LSST). The key challenge is jointly addressing the survey selection function alongside covariant photometric redshift estimation, supernova classification, and distance measurement — problems that are deeply entangled and must be handled consistently to avoid biased cosmological constraints.

Gravitational lensing

Example high dimensional model of a gravitational lens with residuals and data included for reference. Image credit: Legin, R. et al. 2026

I develop fast, differentiable simulation tools for strong gravitational lensing, enabling large-scale inference on lens populations. This work is implemented in the open-source Caustics package.

Astronomical image processing

Example mock supernova data fit using AstroPhot including data, model, and residuals.

I build modular, GPU-accelerated tools for fitting complex astronomical images — handling crowded fields, multi-band/epoch data, and simultaneous PSF characterisation. This work is implemented in AstroPhot.

Astrostatistics and Bayesian methods

Simulation-based inference diagram

My doctoral and postdoctoral work has focused on Bayesian methods for a variety of problems including: galaxy diversity, astronomical image processing, PSF modelling, gravitational lensing, and supernova cosmology.

Collaborators

I am (or have been) affiliated with a number of institutions and collaborations including: