Project 1

---

A musculoskeletal model in OpenSim was trained using Reinforcement Learning with Reference-State Initialization in Python. After 8 hours of training, the agent could walk indefinitely from a chosen starting position. The resulting gait pattern is more natural-looking than other RL agents [bit.ly/33IQ48s]

• Date: May 2021
• Service: Reinforcement Learning

Close

Project 4

---

My team won 1st place at SteelHacks-2021 by making a fake news detector. From user-input URL, the program pulls text data like article title and content, then uses BERT to predict content authenticity [bit.ly/3A3N9nd]

• Date: February 2021
• Service: Natural Language Processing

Close

Project 2

---

As a Computer Vision Programmer with the UWaterloo Aerial Robotics Group, I developed algorithms using OpenCV and TensorFlow for our autonomous aircraft to land, find boxes, and verify their identity via QR codes. The boxes are picked up and delivered to a "medical tent" using GPS guidance [bit.ly/3A6iDJl]

• Date: January 2021
• Service: Machine Learning

Close

Project 3

---

During the MedHacks Hackathon, my team and I created a tumour detection website in Python. We trained a convolutional neural network to identify 7 different types of skin conditions. When the user uploads an image, they're given predictions and directed to learn more about their predicted condition [bit.ly/3rnTwxD]

• Date: September 2020
• Service: Machine Learning

Close

Project 5

---

The No-Touch-Bell is a prototype device which aims to help people with rheumatoid arthritis lift weights. These patients experience pain in their hands when handling heavy weights, but heal faster when they regularly exercise. With this device, the dumbbell is attached to the forearm via Velcro, eliminating the need for hands. The prototype was modeled in SolidWorks then 3D printed and assembled [bit.ly/3rrpwRs]

• Date: December 2019
• Service: Biomedical Device Design

Close

Project 6

---

The TTC's streetcar was recreated in SolidWorks CAD. Interior measurements were taken from real streetcars over multiple rides, and multiple online images were referenced. The project took 30+ hours to complete due to the large number of detailed parts and complex surfaces, but I learned a lot about how to use SolidWorks tools efficiently [bit.ly/3tClgkP]

• Date: December 2019
• Service: SolidWorks CAD

Close