Stefan De Lasa

Experience

• 

  Western University

  Jul 2017 - Jul 2017

  Research Assistant

  Fluidization is a technique used in different industrial processes to alter the flow of gases and liquids. This technique can also be used to make particles (e.g., sand, biomass) behave like liquids. To understand how different experimental conditions impact results of fluidization, I assisted Dr. Nicolas Torres by: • Developing experiments using fluidized beds, under the supervision of a professor• Collecting and inputting data for subsequent analysis• Following hazardous workplace materials protocols Show less

• 

  York University

  Jul 2018 - Aug 2018

  Research Assistant

  Permeable pavements are a promising method to control stormwater run-off. Dr. Usman Khan’s groupis investigating different ML techniques from mobile device input to categorize different paved surfaces.• Conducted experiments to gather important properties of different paved surfaces.• Analyzed and inputting results into a database for subsequent training by a ML model.

• 

  Independent Electricity System Operator (IESO)

  Jun 2021 - Aug 2021

  Meter Data Management Student

  As the Crown corporation responsible for operating/directing the electricity market in Ontario, the IESOgathers and monitors data from industrial customers throughout the province.• Leveraged my technical knowledge to propose how machine learning could improve existing processes.• Highlighted benefits of supervised learning to detect data anomalies using IESO’s historical datasets.• Worked with peers to review Meter Service Provider data, ensuring correctness of meter billing reports. Show less

• 

  PointClickCare

  May 2022 - Aug 2022

  Software Engineer Co-op

  PCC creates healthcare software to assist vulnerable populations with out-of-hospital care.• Migrated the US “Care Insights” application to Canadian markets. Configured a back-end Spring Boot controller to determine session permissions via API calls.• Developed front-end features, using React and Typescript in Docker environments, to ease editing of patient screening templates. Several users mentioned improved usability from my work.• Extracted and used session analytics to guide workflow improvements. Sent metrics to PCC’s Pendo analytics system for subsequent analysis and template refinement.• Used Cypress and Kotlin to write service/unit level tests to ensure UI and data pipeline integrity.• Used Agile development (10 person team) to organize and execute on team commitments Show less

• 

  York University

  May 2023 - Sept 2023

  Computer Vision Research Assistant

  Prof. James Elder computer vision lab is investigating the relationship between image semantics (classification of objects in image) and computer depth estimation. His team aims to build a geometry and semantic based model to see to what extent depth estimation can be turned into an image segmentation problem. During my time, I:• Formulated various methods to use our model to enhance existing Deep Learning depth estimations.• Used Python to implement and evaluate these methods, demonstrating a near 50% improvement. • Was awarded the Student Choice Best Presentation award at the Lassonde Undergraduate Research Conference. Show less

• 

  CentML

  Sept 2023 - now

  Software Development Intern

  CentML is making machine learning training/inference more affordable and efficient. Currently, I'm working on a remote compilation feature that will enable AI model creators to improve their model's performance, resulting in saved time and costs.