Biography
I am currently a full-time Associate Data Scientist at Amazon Web Services (AWS). I work with customers to research and implement cutting edge generative AI applications and ML solutions to realize desired business outcomes. During my time at Amazon, I have worked with large language, multimodal, and diffusion models, MLOps / LLMOps, geospatial computer vision and data science, as well as time-series forecasting.
Previously, I was a Master’s in Computer Science student at the University of Toronto and a member of the University of Toronto Robotics Institute. I was co-supervised by Prof. Florian Shkurti, as a member of the Robot Vision and Learning (RVL) Lab, and Prof. Igor Gilitschenski, as a member of the Toronto Intelligent Systems Lab (TISL) where I conducted research at the intersection of deep learning and robotics, with a focus on model uncertainty calibration and propagation for autonomous vehicles.
- Generative AI
- Deep Learning
- Computer Vision
- Robotics
-
MSc in Computer Science, 2023
University of Toronto
-
BASc in Engineering Science, 2022
University of Toronto
Publications
Experience
- Researching, implementing, and productionalizing AI/ML solutions for customers worldwide to realize their desired business outcomes; specializing in Generative AI models.
- Worked with Sr. Research Science Manager Rohit Malshe and Sr. AI/ML Scientist Arkajyoti Misra in the Amazon Last Mile Planning Science Team.
- Developed a deep learning-based package volume forecasting pipeline using AWS Forecast to predict Amazon’s future customer demands for delivery station planning, outperforming the in-production forecasting model over 90% of the time.
- The project was completed successfully and a return offer was extended.
- Designed and built an end‐to‐end, pixel‐level mangrove forest classification pipeline that could differentiate between mangrove and non‐mangrove regions in satellite images with >90% accuracy.
- Documented my contributions in a tutorial‐style blog post for the official AWS Blog.
- The project was completed successfully and a return offer was extended.
- Worked with Prof. Florian Shkurti in the Robot Vision and Learning (RVL) Lab, self-funded through an Engineering Science Research Opportunities Program (ESROP) Award.
- Developed a novel CNN-based deep learning architecture for one-shot exemplar-based visual search from image and LiDAR data using PyTorch and OpenCV. Experimentally evaluated performance on the KITTI autonomous driving dataset, outperforming state-of-the-art methods.
- Worked with Prof. Alex Mihailidis and Dr. Aaron Yurkewich in the Intelligent Assistive Technology and Systems Lab on the Hand Extension Robot Orthosis (HERO) Glove.
- Implemented a self-contained gesture recognition algorithm that leveraged support vector machines to learn and detect patients’ specified hand and arm movements from onboard accelerometer data to actuate the glove, using Python and C++.
- Designed and developed motion capture software using a PCB with a microcontroller and accelerometer, using C++.
- Assisted in developing a control algorithm using electromyography data which significantly improved hand function during clinical trials.