Frédérik Paradis
Data Science-Engineer
Quebec, Canada
Summary
👤
Senior🎓
Top SchoolFrédérik Paradis is a Data Science-Engineer with 14 years of experience who combines deep academic training—now a PhD candidate in machine learning at Université Laval—with hands-on production engineering in data platforms and ML pipelines. Currently on Ticketmaster’s Pricing Data Science team, he designs and maintains data marts, training/inference pipelines and IaC-driven deployments using Databricks, PySpark, Kubernetes, Pulumi and Terraform. He previously led development of the Poutyne deep learning framework, contributing to usability, testing and documentation, and has meaningful open-source work in computational geometry (CGAL) and live-training tooling (livelossplot). Comfortable moving between research, library-level code and large-scale production systems, he also emphasizes pragmatic developer practices—linting, formatting and automated testing—to keep models reliable and maintainable. Based in Quebec, he blends a computational-geometry background with ML interpretability research, a combination that informs both algorithmic rigor and production readiness.
14 years of coding experience
7 years of employment as a software developer
PhD candidate, Computer Science, Machine Learning Interpretability of Neural Networks, PhD candidate, Computer Science, Machine Learning Interpretability of Neural Networks at Université LavalTechnique, Computer Software Engineering, Technique, Computer Software Engineering at Cégep de Sainte-FoyMaster's degree, Computer Science, Computational Geometry, Local Routing in Spanners Based on WSPDs, Master's degree, Computer Science, Computational Geometry, Local Routing in Spanners Based on WSPDs at University of Ottawa
French, English
Github Skills (14)
computational-geometry10
algorithms10
pytorch10
jupyter-notebook10
c-language10
cgal10
templater10
template-engine10
geometry10
cprogramming-language10
template-tags10
boost9
deep-learning9
keras7
Programming languages (9)
Github contributions (5)
stared/livelossplot
Jul 2018 - Apr 2022
Live training loss plot in Jupyter Notebook for Keras, PyTorch and others
Role in this project:
ML Engineer
ML EngineerContributions:6 commits, 7 PRs, 5 comments in 3 years 9 months
Contributions summary:Frédérik's primary contribution involved adding and updating an example notebook demonstrating the integration of the `livelossplot` library with PyTorch's Poutyne framework for live loss plotting during model training. The commits include modifications to the notebook to align with the latest Poutyne version, define a model and the plotting callback and adjustments to package names. The user also added a test to verify the integration of `livelossplot` with Poutyne.
keras-visualizationpytorchpythondeep-learningjupyter-notebook
CGAL/cgal
Mar 2016 - Dec 2016
The public CGAL repository, see the README below
Role in this project:Back-end Developer
Back-end DeveloperContributions:18 commits, 6 PRs, 31 comments in 9 months
Contributions summary:Frédérik primarily contributed to the CGAL library, focusing on extending the functionality of cone spanners. Their work involved adding options for constructing half-theta and half-Yao graphs, which required modifying existing code and introducing new parameters. The user also developed an Ipelet for cone spanners, which allows users to visualize and interact with these graphs within the Ipe environment. Furthermore, they improved the code by correcting spelling mistakes and refactoring the enum.
boolean-operationsc-plus-plustessellationtemplate-librarypoint-cloud
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.