Experience
Where I've worked and what I've done
Machine Learning Research Assistant
- Built a financial Reinforcement Learning environment Stable-Baselines3 (SB3) for Rotman’s trading simulator
- Developed a Proximal Policy Optimisation (PPO) agent achieving 93% accuracy in arbitrage strategies.
- Extended to a distributed multi-agent PPO training framework in Ray RLib to analyse liquidity dynamics, agent interactions, and emergent behaviours such as collusion in illiquid markets.
- Leveraged RayTune for hyperparameter optimisation and Weights & Biases for end-to-end experiment tracking
Data Engineering Research Assistant
- Implemented an ETL pipeline to collect, parse, normalize, and store data from external APIs and dynamic web sources
- Leveraged asynchronous network requests (urllib + asyncio) to improve data ingestion throughput by 60%
- Automated daily data collection and processing workflows using GitHub Actions and cron jobs; managed reliable data storage using Dropbox SDK
Software Developer and IT Intern
- Maintained and updated all company websites using WordPress; developed and customized PHP templates, themes, and plugins to enhance functionality and user experience
- Managed IT infrastructure, including VPN, firewall, remote desktops, and Office 365 user management; improved system security and operational efficiency through proactive system management
Senior AI Data Trainer
- Specializing in code review of LLM data
- Performed quality assurance and evaluation of code generation (RLHF) across various programming languages (Python, Java, HTML/CSS, JavaScript, C, SQL, TypeScript), achieving 13.5% performance improvement while reducing operational costs by 50%
- Conducted data curation and annotation tasks for Command R/R+ (104B parameters), collaborating with teams to enhance LLM performance
- Collaborated with the synthetic data team on large-scale common crawl web scraping for LLM training pipelines.
- Reviewed and audited tasks from team members, delivering feedback to maintain data accuracy and consistency
AI Data Trainer
- LLM data evaluation and testing
- Prompt engineering and data curation for LLM training
Tech Stack
Technologies I've been working with
Programming Languages
Frameworks & Libraries
Tools & Databases
Machine Learning
Projects
Some cool things I've built and worked on
Scriptorium
A full-stack collaborative code sharing platform using TypeScript, Next.js, Prisma, and SQLite, featuring syntax highlighting, blog post sharing, and code template forking capabilities. Implemented responsive frontend with React, TypeScript, TailwindCSS and used Docker for secure, isolated deployments.
A11YMOLY
A web accessibility testing platform for 0 Barriers Foundation (nonprofit) using React, MongoDB, and Tailwind CSS that automates WCAG compliance scanning and PDF accessibility testing. Implemented detailed violation reporting and compliance scoring system.
UofT ASA DataFest 2024
Secured 3rd place in University of Toronto's ASA DataFest 2024, delivering a comprehensive student engagement analysis report for CourseKata using Python (pandas, matplotlib) and R.
Meal Master
A meal planning application in Java following SOLID principles and Clean Architecture to generate personalized recipes with features such as calorie tracking, weekly meal scheduling, and grocery list creation.
Grade Tracker
A native macOS application using SwiftUI and Swift that helps students consolidate and track assignment grades in a single platform. Built smart grade calculation features including weighted averages, target grade tracking, and predictive insights for required scores.
Movie Match
A Python software that uses community-sourced movie reviews and graph algorithms with cosine similarity. It uses MovieLens dataset with 100,000 ratings from 600 users across 9,000 movies.
Get in Touch
I'm always open to new opportunities and interesting projects. Feel free to reach out!