Fabrice Harel-Canada

Fabrice Harel-Canada

Computer Science PhD

University of California, Los Angeles


I am a second year computer science PhD student at UCLA where I am advised by Miryung Kim and am a member of both the Software Evolution and Analysis Lab and the Statistical Machine Learning Lab.

My research aims at understanding the intricacies of training and testing ML systems. Previously, this entailed devising novel adversarial attack regularizations and investigating test metrics for deep learning models with respect to defect detection, input naturalness, and output diversity. At the moment, I am working on unifying and expanding data augmentation and metamorphic testing techniques into a broader framework of invariant, sibylvariant, and mixture transformations.


  • PhD | Computer Science, 2023

    University of California, Los Angeles

  • MS | Computer Science, 2019

    University of California, Los Angeles

  • BS | Information Systems, 2012

    University of Florida



Teaching Assistant | Software Engineering [CS 130]


Sep 2020 – Dec 2020 Los Angeles, California
  • Developed teaching materials, activities, and exam questions that reinforced the lectures and gauged conceptual proficiency.

Research Assistant


Jul 2020 – Oct 2020 Los Angeles, California
  • Conducting LDA topic modeling and sentence clause-level sentiment analysis using BERT transformers on millions of Amazon and Glassdoor product / employer reviews.

Graduate Student Researcher


Jun 2019 – Present Los Angeles, California
  • Evaluating the reliability of testing metrics for deep learning
  • Developing extensions to adversarial attack algorithms

Reader | Database Systems [CS 143]


Jan 2019 – Mar 2019 Los Angeles, California
  • Developed test cases for a Spark project in Scala, actively addressed student questions, and assisted students with project work


Axenzi Consultants

Oct 2018 – Present Los Angeles, California
  • Helping clients to plan and implement technology projects - specs, devops, and app / database development

Project Manager + Developer Team Lead

Verox Tech

Jul 2015 – Sep 2018 Los Angeles, California
  • Designed and successfully launched a smarter insurance app by building, motivating, and guiding a team of 23

  • Contributed 100s of check-ins, primarily consisting of C# Web APIs and database tables, triggers, sprocs, and T-SQL scripts

  • Designed and implemented a 3-system legacy data migration

  • Responsible for producing, coordinating and reviewing nearly every aspect of the SDLC:

    • code reviews and code optimization
    • requirements and spec generation
    • database design + SSIS + SSRS
    • technical documentation
    • UI design standards
    • product design
    • server setup / config
    • build and deployment scripts

Business + Data Analyst

People’s Trust Insurance

Dec 2012 – Sep 2018 Deerfield Beach, Florida
  • Go-to data expert for executives to produce hundreds of reports and analyses for pivotal decision making at the executive level
  • Developed numerous front-end / back-end application features that were successfully deployed into production to automate tedious and time-intensive workflows
  • Performed continuous requirements gathering & demos with business stakeholders to support 60+ sprints of development for both applications and business intelligence teams, including the transition to a new policy and claim management system


Predicting Stock Prices with Pyro and GNNs

Learning company relations from stock price movements alone can be extremely profitable.

MODE: Automated Neural Network Debugging Replication Project

Open source implementation based on the original paper by Shiqing Ma et al. with a comparison of newly considered distance functions for selecting next training inputs.








US Politics







  • 420 Westwood Plaza, Los Angeles, CA 90095