2026-11-02 –, Conference Foyer 1
Astrophot is a mature tool for parametric image modelling of astronomical images. One has the flexibility to construct arbitrarily complex models of arbitrarily many objects in a scene over arbitrarily many images at different epochs/wavelengths. This is done through an intuitive object oriented interface, and assisted by a large suite of detailed Jupyter notebook tutorials and documentation describing every model. It operates in PyTorch or JAX allowing for accelerated CPU or GPU processing; further, with a caskade backend it is able to trivially interface with other codes such as caustics to add gravitational lens modelling capabilities. In this talk I will detail some of the applications for which AstroPhot is already in use, as well as ongoing transient scene modelling work for the LSST.
I do research software development that is out of this world! My focus is on principled inference using clean modular code. This has lead me to a number of distinct astronomy fields, including image modeling, strong gravitational lensing, and supernova cosmology. My main projects: AstroPhot, caustics, caskade, and PTED are built to scale to new large survey programs and easy to use!