Jan 2024 – Jun 2024 · 5 mos · Toronto, Ontario, Canada · Hybrid

• Self-Driving Car Team, won 1st place in North America
• Fabricated an Extended Kalman Filter (EKF) in C++ and ROS2 for vehicle pose estimation, so that GPS, IMU, and LiDAR data could be fused in real-time, which means accurate localization, resulting in 35% higher accuracy.
• Extracted and cleaned vehicle blob data from SQLite into Python using GenAI, so that it could be fed into the EKF pipeline, introducing an automated preprocessing system to the team, generating 10x faster data ingestion speed.
• Drove key software initiatives for the autonomy team, with a focus on excelling benchmarks and achieving robust system performance, winning 1st place for the fourth consecutive year at the GM/SAE AutoDrive Challenge.

May 2023 – Aug 2023 · 3 mos · Toronto, Ontario, Canada · Hybrid

• Made banking seamless for 5M+ clients. Seen the ability to schedule your eTransfers? Have you accessed your funds? Both of those were me ✋
• Developed loading screens for scheduled eTransfers in Swift and UIKit, so that users could automate transactions seamlessly at a greater convenience to them, accounting for 26% of all eTransfers within the first year.
• Remodelled legacy components to unify UI design patterns, boosting visual consistency across the app, which means less rework for designers, empowering a 25% faster design-to-dev turnaround.
• Reviewed pull requests, optimized code readability, and converted Objective-C to Swift, so that code follows clean and modern design principles, impacting the developer onboarding experience in a 5% faster dev ramp-up time.
• Adopted Test-Driven Development (TDD) across app modules, so that bugs could be caught early in CI, which means cleaner merge pipelines, remediating user-facing issues for around 500K accounts.
• Activities: iOS Workshop Host, Battlesnake, Case Competition

Sep 2022 – Apr 2023 · 7 mos · Montreal, Quebec, Canada

• iOS, Android, Cloud, DevOps. Mainly mobile, but dabbled with a little bit of everything. An impact on 15M+ and improving test coverage by 15%. I also spoke in front of 1000+ student at a career fair, yay.
• Spearheaded POCs and shipped cross-platform features in Swift and Kotlin, so that core ideas like dark mode and name-based greetings could be validated early, giving rise to a 2x faster buy-in from stakeholders.
• Refactored iOS UI tests to improve coverage by 15%, so that regressions could be caught earlier, which means higher release confidence, contributing to smoother QA sign-offs each sprint.
• Resolved Android and iOS layout bugs like Apple Wallet sizing and sticky footers, facilitating a streamlined visual experience across devices, affecting 15M+ clients and a 18% reduction in customer complaints.
• Joined cloud team to present AWS tagging improvements, making sure that app traffic could be properly bucketed, which means easier billing and audits, culminating in 100% tag-coverage across services.
• Activities: Cloud Engineering Group, Case Competition, Sun Life Student Influencer Committee, Panelist at the YNCN Career Fair Summit on behalf of Sun Life and spoke in front of 1000+ students

Sep 2021 – Jun 2022 · 9 mos · Toronto, Ontario, Canada

• Lots of testing, and lots of 1st places, whoop whoop
• Leveraged MathWorks Simulink and ROS2 to design autonomous vehicle test scenarios, so that safety features could be validated, which means reduced on-road testing risk, resulting in 5x fewer field bugs.
• Built publisher, subscriber, and routing models in MATLAB to simulate traffic light systems, so that the Mathworks Challenge needs were met, which means accurate V2X testing, leading to 1st place.

May 2021 – Aug 2021 · 3 mos · Houston, Texas, United States

• Where I learned that the book Getting Things Done can be applied in a work setting, this was mind-blowing. Oh and made a SDK that reduced customer onboarding time by 50%, that's a pretty big win.
• Architected JavaScript (JS) SDK using gRPC, protobuf, and OOP concepts, addressing 80% of active JS customers to smoothly integrate the Kurtosis software onto their systems, optimizing the customer onboarding time by 50%.
• Migrated 25% of the Go backend codebase to scalable Typescript modules, allowing clients to consume modern SDKs and establishing reusable tooling across teams, in the form of 4 published npm packages.
• Diagnosed multithreading Golang functions to shut down running Docker containers and implemented in CircleCI, so that system resource leaks could be prevented, creating a 3x improvement in computer speed and storage.

Jan 2021 – Apr 2021 · 3 mos · Vancouver, British Columbia, Canada

• Hired to write math problems, somehow built a full-stack web app instead
• Allocated 80+ hours to create mathematical problems using the Hypatia software & developed full-stack web application in an agile team of 5 with CheckMath’s API using Django and React which saves 25% of time for student users in organizing and discovering weaknesses for a course (during the previous fall term).

May 2020 – Sep 2020 · 4 mos · Toronto, Ontario, Canada

• Data, NLP, and Project Management. Papers accepted into the SIGCSE Technical Symposium and CHI Parallel Presentation.
• Research: Support student learning behaviour through homework reminder emails.
• Applied NLP techniques like frequency analysis and stopword filtering, so that language patterns could be extracted, which means faster iteration cycles, contributing to a 70% faster insight discovery.
• Integrated a spellchecking module into preprocessing pipeline, so that participant data could be cleaned reliably, which means less manual review, resulting in 2x increase in label accuracy.
• Optimized project management processes (sending emails, managing meetings, formulating plans and improving communications channels), reducing training time for new lab members by 30%.
• Conducted literature review and survey coding for research papers, that ended up receiving an acceptance into the SIGCSE Technical Symposium and CHI Parallel Presentation.

Jan 2020 – Apr 2021 · 1 yr 3 mos · Toronto, Ontario, Canada

• Teaching drones how to dodge objects
• Piloted OpenCV pipeline to segment images into patches for neural net input, so that small targets could be detected at range, which means finer spatial resolution, resulting in 4x detection recall.
• Built drone target detection with YOLOv3, gRPC, and PyTorch in a team of 4, so that real-time threats could be classified, which means lower false positives, resulting in 90%+ detection accuracy.
• Studied ML foundations (ex. training, classification, CNN) and research papers, so that I could contribute to PyTorch model tuning, which means better architectural decisions, resulting in 2x faster convergence.