The AMBIENT Project Team Supervised by Babak Taati, University of Toronto

• Built an end-to-end pipeline that automates processing and running machine learning models on raw input videos, and extracts key 3D joint locations to compute gait features and determine fall risk for patients with dementia.

• Assessed the quality of newly collected data and provided insights and recommendations to data collectors in order to maximize the quality of the data collected.

• Performed a quantitative and visual analysis of the performance of the 3D model compared to previous 2D pose analysis models by contrasting the gait feature values and writing scripts that help visualize specific features such as foot contacts.

Led a group of 16 students responsible for fielding 26 teams for a machine learning competition from North American universities such as CMU, Cornell, and Yale, and helped raise over $100,000 by contacting sponsors such as Google, IBM, and the Vector Institute while maintaining relationships with them as a club.

Involved in organizing a speaker series initiative aimed to enrich undergraduate students interested in artificial intelligence.