• Generating an AI-Automation playbook, which includes a deep analysis of current business processes and opportunities for automation to enhance organizational efficiency without increasing head count.
• Creating applications and KPIs to automate metrics, data analysis, and automating business processes.

• Conducted extensive data analysis and categorized large datasets, identifying key factors and trends.
Internship
• Produced a report regarding various parts of machinery at Honda Alliston plant, including clustering of categories
• Analyzed a set of materials to recommend cost-effective substitutes, reducing expenses.

• Drastically reduced vendor communication time by implementing automated solutions.
• Achieved significant cost savings and process efficiencies by optimizing barrel cutting operations.
• Developed document generation tools in collaboration with the Engineering department, automating complex tasks and driving substantial annual savings.
• Streamlined the COGS calculation process through automation, leading to enhanced accuracy and reduced manual effort.
• Generated substantial annual savings and reduced labor hours across various departments through the implementation of advanced automation solutions.

• Collaborated closely with cross-functional teams to design and implement software systems that address complex challenges, ensuring alignment with project goals and industry standards.
• Led and contributed to high-impact research and development projects for organizations, including Honda, Hörmann, and XR-AI, focusing on creating scalable software solutions and advanced data analysis methodologies.

• Led a team of four developers, coordinating efforts to design, develop, and deploy a robust platform that fosters a community-driven approach to software debugging.
• Collaborated with industry mentors to integrate product and business development insights, resulting in a platform that not only meets technical requirements but also aligns with market needs and business goals.
• Managed the project lifecycle from inception to deployment, including planning, resource allocation, timeline management, and risk mitigation, ensuring the successful delivery of a high-quality platform.
• Facilitated regular team meetings, progress updates, and mentor consultations, maintaining clear communication channels and fostering a collaborative work environment.

• Successfully reduced paramedic data response times through the development of an innovative Data Analysis PowerApp.
• Decreased data request turnaround time with the creation of a dynamic Power BI KPI dashboard.
• Improved emergency response time accuracy through the design and implementation of a data pipeline integration.
• Collaborated with the IT team to test and validate the integration of Oracle PeopleSoft ERP's Payroll Processing Module with the HR module.

A smart, automated solution for tracking facility capacity in real-time, predicting future usage, and enhancing the recreational experience for Orillia citizens.
Key Features
👁️
AI-powered Video People Counter
Advanced computer vision technology accurately counts facility visitors in real-time, eliminating manual counting errors and providing precise capacity data.

📊
Future Capacity Prediction
AI-driven predictive analytics forecasts facility usage patterns based on historical data, weather conditions, and special events, helping visitors plan their visits optimally.

🤖
AI Phone Agent & Chatbot
Intelligent virtual assistants provide instant capacity information via phone or online chat, answering visitor questions and reducing staff workload.

AI-powered legal assistance to help you start your case with confidence.

Project Abel is an AI-driven legal assistant designed to help users assess and structure their case before reaching out to legal professionals. We aim to make legal guidance more accessible and empower individuals to navigate complex legal situations with confidence.

Our platform serves as a bridge between individuals seeking legal assistance and the professional legal counsel they need, helping to prepare and organize information for more effective legal consultations.

Project Orbit is an innovative interactive orrery that takes you on a journey through the solar system and beyond. Not only does it offer a web-based interactive orrery, but it also integrates AI, enabling you to chat via text and converse via voice-chat with a virtual assistant for an engaging experience. We’ve gamified the experience by adding spaceship controls, allowing users to explore the galaxy in a game-like environment. The project also features a Keplerian orbital propagator, which simulates planetary motion using real NASA datasets.

Winner of NASA "Innovation" award, 1st place + Global Nominee for 2024 NASA Global Space Apps Challenge.

Featured in CTV News interview: https://www.ctvnews.ca/barrie/article/nasa-challenge-wraps-up-at-lu-orillia/

Hosted Version: https://project-orbit.vercel.app/

This paper investigates the use of machine learning regression models to predict stock market closing (close) prices using large amounts of stock market data from S&P 500 companies. The prices.csv dataset contains over 850,000 rows of stock price data spanning from 2010 to 2016, with features such as opening price, daily highs and lows, trading volume, and closing price. Through feature engineering new metrics were introduced such as average trading price (avgPrice), price range (priceRange), and volatility index (volatilityIndex) to enhance predictive accuracy, with both normalized and non-normalized versions of all numeric columns/features.
Four regression models were used to train and evaluate the model in terms of its successful prediction of the closing prices of stocks. These 4 regression models are: Linear Regression, Random Forest Regressor, K-Nearest Neighbors (KNN), and Decision Tree Regressor. The data split consisted of 80% of the data being used to train the model whilst 20% remained for testing. Performance was assessed using Mean Absolute Error (MAE), Mean Squared Error (MSE), and R-squared (R²) metrics. Models trained on normalized features consistently outperformed those trained on non-normalized data, achieving lower MAE and MSE values. This is due to non-normalized data being prone to being skewed by larger scale features, such as trading volume (volume). Linear Regression emerged as the most accurate model, followed by Random Forest and KNN, while Decision Trees offered interpretability but slightly higher errors.
Results highlight the critical role of feature preprocessing, particularly normalization, in improving model performance. Near-perfect R² values (0.9999) across all models demonstrate the strength of the selected features in capturing variability in the target variable. This study shows the potential of machine learning for financial forecasting and highlights areas for improvement, such as incorporating external factors, time-series features, and expanding dataset scope. These findings pave the way for accessible, AI-driven tools for everyday traders and analysts, enhancing decision-making in dynamic financial markets and leveling the playing field for all.