- Worked on C++ firmware and supporting scripts for the Tahoe DSP, which powers pluggable optical transceivers- Developed a system for applying fractional delays and gains to FIR filter taps to test impairment compensation algorithms- Used MATLAB to process factory calibration data into SQLite databases, ensuring backwards compatibility for older formats- Implemented algorithms such as DFTs in C++20 with the help of lambdas, templates, OOP, and the standard library- Wrote unit tests for all pull requests, which were further reviewed by senior engineers to enforce good practice and correctness

Member of Waterloo Formula Electric, which competes in Formula SAE:- Wrote C firmware for I2C communication between BMI088 IMU and custom STM32-based telematics control unit- Prototyped CAN message logging through SDIO to a microSD card- Added circuitry and firmware to power distribution unit's HIL testing board to imitate DC-DC power supply toggling- Fixed dashboard button detection and double-click issues in dashboard control unit firmware and embedded Debian UI scripts- Implemented APPS/brake pedal plausibility check with FreeRTOS to ensure safety in case of accelerator pedal failure

- Led backend development of generative AI product from inception to prototype with Flask, Selenium, and fine-tuned LLMs- Reduced runtime of newscast video processing pipeline on distributed AWS EKS cluster by ~26%- Automated code deployment, data preprocessing, and result validation using Bash and Python scripts- Created and optimized video encoding, object detection, and text detection algorithms using FFmpeg, OpenCV, and Torch- Containerized algorithms with Docker to facilitate debugging, cloud deployment, and reuse within pipelines

- Led the design, manufacturing, and programming of robots for the annual FIRST Tech Challenge competition, winning the Inspire Award at the 2023 Ontario Provincial Championships (and competing at the FIRST World Championships in Houston) for demonstrating excellence in every aspect of the competition: robot design, robot performance, documentation, and team coherence- Guided a team of 12 members through manufacturing and testing 6 robot revisions over 13 months with a combination of off-the-shelf, 3D printed, and custom-machined aluminum, polycarbonate, and steel parts- Used Onshape to create detailed models of our robot throughout its development, and used Java to code autonomous and driver-controlled programs for the competition, implementing PID control, I2C sensing, and OpenCV/TensorFlow computer visionI may have graduated, but I am still extensively mentoring our talented team!