Zarreen Reza
I am a Senior Applied AI Research Scientist with over five years of experience across diverse industries. My background combines deep research and practical ML expertise, spanning various domains.
I earned my M.Sc. in Computer Science focusing on AI from the University of Windsor. Since then, I have held roles as a Data Scientist at Thales for over two years, an AI Research Scientist at Volta Charging Inc. (an EV charging infrastructure company in San Francisco) for more than a year, and currently, I serve as a Senior Applied AI Research Scientist at Jacobb.ai, a non-profit applied AI research center in Montreal, Quebec.
My current research work focuses on the privacy, fairness, robustness, interpretability, and scalability of LLMs and LLM systems. I have extensive research and industry experience in deep learning, LLMs, and privacy-preserving ML, applying these techniques to human-centered applications such as education, healthcare, and more. Additionally, I am highly interested in doing research on the reasoning capabilities of LLMs.
Research
(*equal contribution, †alphabetical / random authorship)
Low-rank finetuning for LLMs: A fairness perspective
S. Das, M. Romanelli, C. Tran, Z. Reza, B. Kailkhura, F. Fioretto
Preprint, 2024.
Automated identification of sea pens using OpenCV and machine learning
Z. Reza† and 9 others.
Data Study Group, The Alan Turing Institute, 2023.
Published as the final report of the Data Study Group at The Alan Turing Institute in London, UK, following a week-long intensive hackathon.
PySyft: A Library for Easy Federated Learning
A. Ziller†, A. Trask†, A. Lopardo†, B. Szymkow†, B. Wagner†, E. Bluemke†, J. M. Nounahon†, J. Passerat-Palmbach†, K. Prakash†, N. Rose†, T. Ryffel†, Z. Reza†, G. Kaissis†
Springer International Publishing, 2021.
Published as a book chapter in Federated Learning Systems: Towards Next-Generation AI.
Real-time Automated Weld Quality Analysis From Ultrasonic B-scan Using Deep Learning
Reza, Zarreen Naowal
Master’s dissertation, 2019. Nominated for Governor General’s Gold Medal.
Supervised by Dr. Roman Maev and Dr. Dan Wu.
Preliminary work for Patent US20230228716A1 ([0140], [0141], [0142], [0143]).
Detecting jute plant disease using image processing and machine learning
Zarreen N. Reza, Faiza Nuzhat, Nuzhat Ashraf Mahsa, Md. Haider Ali.
ICEEICT, 2016.
Orally presented in the 3rd International Conference on Electrical Engineering and Information Communication Technology (ICEEICT).
Authenticity Analysis of Opinion and Claims made online
Zarreen N. Reza, Ruturaj R. Raval, Kaival Patel.
3rd Winner, Thales Student Innovation Championship in AI, 2018.
Ranked 3rd out of 52 student teams from universities across Canada by developing an end-to-end AI solution for fighting misinformation in online news.
Visual Relationship Detection using Deep Learning for Vision and NLP
Zarreen N. Reza†, Ashraf Neisari†, Ruturaj R. Raval†, Zeeshan Mansoor†.
Course final project, University of Windsor, 2018.
Finding Relevant Biomarkers for Prostate Cancer using Feature Selection and Dimensionality Reduction
Zarreen N. Reza, Susha Suresh.
Course final project, University of Windsor, 2018.
A Comprehensive Study of Unsupervised Feature Learning using Autoencoders
Zarreen N. Reza.
Course final project, University of Windsor, 2019.
Cancer Progression Prediction from Breast Cancer Gene Expression Data using Machine Learning
Zarreen N. Reza, Rajasi Upadhyay, Sumit Khairnar.
Course final project, University of Windsor, 2018.
Experience
Senior Applied AI Research Scientist, 2023 — Present
JACOBB – Applied Artificial Intelligence Center
Montreal, Canada
Co-Founder and Lead AI Scientist, 2021 — Present
resolutio.ai
Ottawa, Canada
Privacy-Preserving AI Researcher, open-source contributor, 2020 — Present
OpenMined
Remote, USA
AI Research Scientist, 2021 — 2023
Volta Charging, later acquired by Shell
Montreal, Canada
Data Scientist, 2019 — 2021
Thales Canada (Guavus)
Montreal, Canada
Research Assistant, 2017 — 2019
Institute of Diagnostic Imaging and Research (IDIR)
Windsor, Canada
Teaching
- Course instructor and content creator of the OpenMined Private AI Series on Federated Learning using PyTorch and PySyft, funded by Meta AI, PyTorch, and the University of Oxford, with a focus on hands-on coding for privacy-preserving AI. 2021.
- Tutor of COMP-8547 (formerly 60-654) Advanced Computing Concepts at University of Windsor. 2018 – 2019.
- Graduate Teaching Assistant at the University of Windsor. 2017 – 2019.
- Teaching Assistant at BRAC University. 2013 – 2016.
Talks
- Career Talks to Guide You Through Tech (AI)
Women Who Code Career Cruise Series. October 2022. Invited. - Mistakes to Avoid as an AI Practitioner in the Industry
Women Who Code CONNECT Recharge. June 2022. Invited. - End-to-end orchestration of machine learning models from PoC to deployment
Data Stack Summit. April 2022. Invited. - End-to-end ML Orchestration for Private Data
Women in Data Science (WiDS) Vancouver Annual Conference. March 2022. Invited. - The Role of Public Speaking in Career Growth
Women Who Code CONNECT Forward. November 2021. Invited. - Privacy-Preserving AI – Perform Data Science on Data You Cannot See
Women Who Code CONNECT Reimagine. June 2021. Invited. - Climbing up the Ladder Through Open-Source Contribution
Panjab University Software Freedom Day. September 2021. Invited. - Private Medical Deep Learning with Federated Learning and Differential Privacy
OpenMined PriCon Privacy Conference. September 2020. - Convolutional Neural Network for Speech Recognition and Diarization
Intel Edge AI Scholarship Challenge Workshop. February 2020. - Weld Quality Control using Artificial Intelligence
3 Minute Thesis Competition. March 2019.
Services
- Leadership Fellow, Women Who Code, Aug. 2021 – Sep. 2022.
- Led the Data Science Track consisting of 4500+ members in organizing free events, and workshops to help women excel in STEM roles.
- Led a team of 50+ volunteers in program designing, event planning, and building technical content for 150+ events including webinars, workshops, hands-on coding tutorials, career growth, etc.
- Honed public speaking skills through speaking at 40+ events including annual summits and conferences.
- Organizer
- Organizer of WWCode Data Science Career Mentorship Program 2022, a four part workshop series attended by 340+ audiences.
- Organizer of WWCode Machine Learning Study Group 2022, an eight part webinar series on beginner to advanced level ML concepts attended by 600+ audiences.
- Co-organized BlockDataPy Tech Summit 2022, a one-day summit consisting of talks in blockchain, data science, and Python.
- Co-organized WWCode Hackathon for Social Good 2022 attended by 200+ participants.
- Judge and Reviewer
- Reviewer, Causality and Large Models (CaLM) Workshop at NeurIPS 2024.
- Judge and project reviewer of TigerGraph’s Graph for All Million Dollar Challenge in March 2022.
- Judge at Windsor Regional Science, Technology, and Engineering Fair (WRSTEF) in March 2019.
- Mentorship
- Mentor at OpenMined Beta Bootcamp’s Tech Team, a bootcamp for Privacy-Preserving Machine Learning (PPML) from July 2020 to August 2020.
- Mentor at Technovation Challenge, the world’s largest tech entrepreneurship program for girls from January 2018 to August 2018.
Mentorship
- JACOBB: Led 8-member AI team, planning and coordinating research work as sprints, ensuring timely milestone delivery.
- JACOBB: Mentored five junior AI researchers, focusing on code quality and standardization.
- OpenMined: Mentored 25 participants of the OpenMined Beta Tech Bootcamp on Privacy-Preserving ML.
- Guavus: Mentored two data scientist interns and collaborated with researchers, engineers, and the customer success team on AI model integration.
- IDIR: Trained three undergrad CS students in Python Programming and Machine Learning to succeed in their Co-op term.
Awards and Achievements
- AI Innovator of the Year, 2023
Awarded as a finalist by Women in AI Awards for contributions to AI innovation, leadership, and advancing the AI ecosystem in North America. - Governor General’s Gold Medal, 2019
My master’s thesis was nominated by the thesis committee at UWindsor for this highly prestigious award. - Director’s Honour Roll, 2019
Awarded for academic excellence from the Director, School of Computer Science, University of Windsor. - Going Beyond and Above in Research Award, 2019
Nominated for outstanding research performance and significant contributions at both national and international levels, as recognized by my master’s thesis supervisors Dr. Roman Maev and Dr. Dan Wu at UWindsor. - Ambassador Award, 2019
Nominated by Dr. Ziad Kobti, Director of the School of Computer Science at UWindsor, for raising the university’s profile by winning 3rd prize out of 52 teams in a national AI competition. - NSERC CREATE Highly Qualified Personnel (HQP) Award, 2018
Awarded to highly qualified graduate students to train them in industrial research with a strong involvement with the Canadian industry. Received upon nomination by my master’s supervisor, Dr. Roman Maev at UWindsor. - Vice Chancellor’s Award for Academic Result, 2012-2015
Awarded 3 times for receiving the highest GPA in BRAC University. - Vice Chancellor’s Award for Co-curricular Activities, 2012
Awarded for outstanding all-round performance in the Residential Semester at Savar Campus of BRAC University. - Winner, Encode Club x Polygon Hackathon, 2022
Developed a blockchain-based solution to tackle NFT duplication and theft. - 3rd Winner, Encode Club x Filecoin NFT Hack, 2021
Developed a blockchain-based platform for artists to detect and seek justice on their stolen artwork through decentralized dispute resolution. - Finalist, Qiskit Hackathon Europe: Research Study Group, 2021
Implemented Quantum Support Vector Machine (QSVM) for predicting astronomical events from limited amount of data, outperforming classical SVM in terms of accuracy. - OpenMined Featured Contributor, 2020 and 2023
Recognized twice as an outstanding contributor to one of the largest open-source communities in AI. - 3rd Winner, Thales Student Innovation Championship in AI, 2018
Ranked 3rd out of 52 student teams from universities across Canada by developing an end-to-end AI solution for fighting misinformation in online news.
Certificates
- OpenMined Padawan Program Graduate, 2022
Completed a rigorous six-week mentorship program focused on contributing to PySyft and open-source projects in the field of Privacy-Enhancing Technologies (PETs). Contributed to OpenMined’s open-source repositories through coding, documentation and technical challenges. - IBM Certified Associate Developer – Quantum Computation, 2021
Passed the exam for demonstrating advanced knowledge of quantum computing concepts, utilizing Qiskit SDK with Python to develop and execute programs on IBM Quantum computers and simulators. - Quantum Computing Certificate, The Coding School (Sponsored by IBM Quantum), 2020-2021
Completed 8 months long program garnering in-depth skills on Mathematics for Quantum Computing, Quantum Mechanics, Quantum Algorithms, Quantum Applications, Quantum Hardware, Coding in Quantum Computer, Quantum Cryptography, and more. - Quantum Excellence Certificate, Qiskit Global Summer School by IBM Quantum, 2020
Passed the course with hands-on experience on quantum computing, including designing and simulating a Variational quantum eigensolver (VQE) algorithm for calculating the ground state energy of the LiH molecule.