Mathew Rogers
I am a data scientist in Montreal. Over the last few years, I have worked in machine learning, streaming data, and computer vision. My various roles have all included a significant software engineering component. My PhD is in number theory with a focus on conjectures relating to elliptic curves. Feel free to reach out, I'm always looking for a challenge!
Experience
Shopify
Senior Data Scientist • 2021 — 2022
Implemented real time metrics for the Shopify platform. Joined and aggregated multiple Kafka streams using Scala and Apache Flink. Utilized various streaming data algorithms and test-driven development.
Via Science
Deep Learning Principal • 2018 — 2021
Data Scientist • 2017 — 2018
Projects involving machine learning, data science, and time-series forecasting in the energy sector.
- Implemented machine learning based data ingestion pipelines, in order to ingest heterogeneous smart meter data into a standardized schema.
- Fine-tuned and dockerized BERT based models for NLP related tasks.
- Contributed to federated learning, homomorphic encryption, and differential privacy projects.
- Prepared project proposals for clients.
Clemex Technologies
Machine Learning Scientist • 2015 — 2017
Extensive use of convolutional neural networks, OpenCV, and traditional computer vision while working in the microscope industry.
- Implemented and fine-tuned state-of-the-art convnets for image classification and segmentation.
- Created numerous datasets ranging from thousands to millions of images from scratch.
- Implemented and queried relational databases for handling training data.
MILA, Université de Montréal
Graduate Student • 2014 — 2015
Experience as a student working in the LISA lab (now MILA) developing Theano and Pylearn2. Completed a full set of graduate-level computer science courses.
Université de Montréal and CRM
Postdoctoral fellow • 2011 — 2013
Researcher at the Université de Montréal and the Centre de Recherches Mathématiques (CRM). Teaching duties at McGill University and Concordia University.
University of Illinois, Urbana-Champaign
NSF Postdoctoral Fellow • 2008 — 2011
Research position that successfully solved a 30 year old conjecture in pure mathematics using unexpected methods.
- Jointly developed the Rogers-Zudilin method for expressing L-values of elliptic curves as special values of regulators.
- Solved the Deninger-Boyd Mahler measure conjecture for conductor 15 elliptic curves.
- The approach is currently state-of-the-art in the area.
Max Planck Institute for Mathematics, Bonn
Visitor • 2009, 2010
Visiting researcher at the Max Planck Institute for one year.
Projects
Patents
• 2017 — 2021
- Secure data processing, MD Rogers, KC Cheung, J Taylor, US10972261
- Secure data processing, KC Cheung, J Taylor, M Rogers, C Gounden, US11038683
- Multiple data source secure data processing, KC Cheung, M Rogers, J Taylor, US11489659
Published Papers in Mathematics
• 2005 — 2013
Published 19 papers while working as a mathematician. A full list is available on Orcid or Google Scholar. Preprints are available on the Arxives.
Blog
• Check Back Soon...
Education
University of British Columbia
PhD, Mathematics • 2008
Graduated with a focus in analytic number theory and experimental mathematics. My work focused on conjectures involving L-values of elliptic curves. Math department nominee for Canada wide thesis award.
Additional Links
Social Links
- Github: https://github.com/matrogers
- Twitter: http://twitter.com/matrogersmtl
- LinkedIn: https://www.linkedin.com/in/matrogers31