Machine Learning Researcher at BMO Lab, University of Toronto

Built a real-time motion ingestion and streaming pipeline in Python to feed motion-capture data into a humanoid simulation system, handling latency buffering and sequence delivery for synchronized motion playback in a live simulation. Improved simulation stability and training data quality by refining body fitting, temporal transitions, joint smoothing, and denoising, reducing noise and discontinuities in motion sequences and lowering humanoid fall-down failure rate by 26%. Engineered real-time contact-force and per-joint torque extraction from the simulator API, streaming outputs via OSC/UDP for live monitoring, visualization, and dataset generation to support downstream model training.

Senior Consultant, Application Development at CIBC

Engineered production-grade backend services in Python and C#/.NET for low-latency, event-driven algorithmic trading workflows, integrating IBM MQ and Solace to support time-sensitive trading and operational processes at production. Built and supported full-stack web applications with React, JavaScript, backend services and REST APIs, delivering tools for real-time monitoring, analytics, and report scheduling for applications supporting $500M–$1B in order volume. Led technical consultation with trading desks to translate trading strategies and reporting requirements into actionable engineering plans, while supporting reliable software delivery through Azure DevOps, CI/CD, and release pipelines.