I'm a Computer Science graduate from the University of Toronto with concentrations in Machine Learning, Computer Vision, and AI. I hold a 3.55 GPA, made the Dean's List, and received the University of Toronto Scholars Award.
Most recently I spent over a year at AMD as a Software and Testing Engineer, building ML-powered automation tools and having my Test Workflow Optimizer showcased as a finalist at AMD's internal innovation event. Before that, I was at Signifi Solutions building full-stack integrations and real-time data pipelines.
My proudest achievement: a drone-based soccer analysis system using computer vision and RNN trajectory prediction that was selected for presentation at MLSE (Maple Leaf Sports & Entertainment).
Built an ML pipeline using TF-IDF, stemming, K-Means, and DBSCAN to cluster QA test cases, reducing setup time by ~15% and improving execution efficiency. Delivered via Python + custom Tkinter GUI; showcased as a finalist at AMD's internal innovation event.
Developed a Python–PostgreSQL integration with the Odoo API to automate ticket data retrieval and analytics, enabling real-time Power BI reporting for the operations team.
Built a drone-based computer vision system with multi-object tracking and RNN trajectory prediction to deliver real-time player/ball analytics — possession tracking, heatmaps, and tactical phase detection using the SoccerTrack dataset.
Developed and benchmarked detection/tracking pipelines (MOTA, ADE/FDE) alongside a user-facing analytics interface, enabling affordable professional-level tactical insights for lower-division teams. Selected to present at MLSE (Maple Leaf Sports & Entertainment) Innovation Day.
import torch
from tracking import MultiObjectTracker
from models import TrajectoryRNN
# Drone-based CV analysis pipeline
class SoccerAnalyzer:
def __init__(self):
self.tracker = MultiObjectTracker()
self.rnn = TrajectoryRNN()
def analyze(self, frame):
# Detect players + ball
detections = self.tracker.detect(frame)
# Predict trajectories (RNN)
traj = self.rnn.predict(detections)
return {
"possession": calc_possession(detections),
"heatmap": gen_heatmap(traj),
"tactics": analyze_tactics(traj),
"mota_score": self.tracker.eval_mota()
}
Built an ML-powered tool at AMD to cluster QA test cases using TF-IDF, stemming, K-Means, and DBSCAN — reducing setup time by ~15% and improving execution efficiency. Integrated with internal APIs via a custom Tkinter GUI.
City-wide bike rental app with Flutter/Dart frontend and Python backend. Features scalable database architecture, Stripe payment integration, and real-time ride tracking. Built for production readiness.
Lightweight HTTP web server in C++ built from scratch using sockets. Handles TCP connections, static file serving, multithreaded client handling, MIME type resolution, and robust error handling.
Python–PostgreSQL integration with the Odoo API to automate ticket data retrieval and real-time analytics. Enabled live Power BI reporting and significantly improved data management efficiency.
Peer-to-peer direct messaging system in C using advanced socket programming for reliable, real-time communication between users over a network.
Recursively compresses and decompresses files of all types (images, audio, text) by converting them into Huffman trees where each node represents a data point.
Probabilistic classifier that predicts the most likely outcome given a factor table — applied to medical condition likelihood prediction using Big Data strategies.
Group project implementing the word-finding game Boggle with a full GUI in Java, applying Agile/Scrum methodology and collaborative Git workflows.
Automated data transfer from Freshdesk API to Excel using VBA at Signifi, creating automatic backups and improving ticket system data visibility.
Determines monthly bills for telephone plan subscribers with user-defined filters for tracking specific customers and calls, visualized through a GUI.
Centralized internal web portal at Signifi for scheduling, uploading, and sending company-wide newsletters, built solo from scratch.
Fully functional Tetris implementation with a working GUI built in Java, handling game logic, rendering, and user input.
Unix/Bash program that copies files between directories while intelligently filtering for duplicates, using shell scripting and file system operations.
I'm currently open to new opportunities. Whether you have a question, a project idea, or just want to connect — feel free to reach out!
Download Resume