dfreproject is a high-performance Python package for astronomical image reprojection, optimized for both GPU and CPU execution. Developed for the Dragonfly Telephoto Array, it addresses the computational bottleneck of aligning and stacking astronomical images by providing GPU-accelerated coordinate transformations and interpolation.

The package achieves up to 20X speedup on GPUs and 10X on CPUs compared to traditional methods, making it essential for processing large volumes of data from modern wide-field imaging surveys. Published in the Journal of Open Source Software (JOSS), dfreproject follows FITS and SIP standards for maximum compatibility.

LUCI is a comprehensive line fitting pipeline built specifically for the SITELLE IFU at the Canada-France-Hawai'i Telescope. It natively integrates cutting-edge machine learning techniques to provide accurate and efficient spectral analysis.

The software features extensive documentation, including detailed examples, explanations of the underlying algorithms, and complete API documentation. LUCI represents a significant advancement in automated spectroscopy analysis for large-scale IFU surveys.

Pamplemousse is a suite of advanced machine learning applications designed specifically for the SITELLE imaging Fourier transform spectrometer at CFHT. Building on the LUCI pipeline, Pamplemousse provides cutting-edge tools for emission line classification, kinematic analysis, and spectral decomposition using state-of-the-art deep learning architectures.

The package implements convolutional neural networks for automated HII region classification, mixture density networks for multi-component line fitting, and specialized algorithms for handling the unique characteristics of SITELLE data. These tools have been instrumental in several publications and are used by the SIGNALS collaboration.

This project introduces a novel recurrent neural network architecture for deconvolving X-ray spectra from galaxy clusters. The Recurrent Inference Machine (RIM) addresses the fundamental challenge of extracting accurate temperature and abundance measurements from complex, multi-temperature X-ray emission by learning to iteratively refine spectral decompositions.

Presented at the NeurIPS 2023 Machine Learning and the Physical Sciences workshop, this work demonstrates how physics-informed deep learning can solve inverse problems in astrophysics. The method achieves superior performance compared to traditional fitting techniques, particularly in handling overlapping spectral components and low signal-to-noise data.

Valence is a sophisticated web application built with Django and jQuery that enables users to create cognitive-affective maps—visual representations of complex mental models and emotional associations. The platform provides researchers with powerful tools for qualitative and quantitative analysis of these cognitive structures.

Originally developed for the Basilie School at the University of Waterloo and later enhanced through a contract with the University of Freiburg, Valence has been instrumental in several published research studies on cognition and decision-making.

The LEMUR (Led by Extreme Multi-wavelength Understanding Research) X-ray Galaxy Cluster Archive provides open access to a comprehensive catalog of X-ray observations of galaxy clusters. This project combines automated data reduction pipelines with systematic analysis to create a valuable resource for the astrophysics community.

The archive serves as a foundation for large-scale statistical studies of galaxy cluster properties, evolution, and cosmological implications. It showcases advanced X-ray data analysis techniques and effective presentation of complex scientific data.

Cadena is an innovative sharing economy platform that enables users to buy and sell books without the typical 20% platform fees charged by competitors. Built as a full-stack Django application, Cadena operates on a donation-based model, encouraging users to contribute voluntarily after successful transactions.

This was my first major web development project, developed in collaboration with Christian Thibeault. Through this project, I gained expertise in Django, built a complex authentication system, payment processing, and learned the fundamentals of creating a production-ready web application ecosystem.

AstronomyTools is a comprehensive collection of Python utilities and analysis scripts for astronomical data processing. Born from a commitment to open science, this repository shares battle-tested code developed throughout my research career, making advanced analysis techniques accessible to the broader astronomy community.

The toolkit includes modules for X-ray spectroscopy, optical data reduction, machine learning applications in astronomy, and various visualization utilities. All tools are well-documented and designed to be easily integrated into existing analysis pipelines.