As a Machine Learning Engineer Intern at Blackberry from January 2024 to April 2024, I played a pivotal role in enhancing the company's alert system. Leveraging my expertise, I developed novel machine learning models capable of identifying false positives within vast databases containing over 450,000,000 alerts. By implementing Decision Tree and statistical models using Python, Databricks, and Pyspark, I successfully reduced alert fatigue by more than 35%. Additionally, I harnessed Lasso Regression and Principal Component Analysis (PCA) techniques to pinpoint the four most influential features within the datasets. My work extended beyond model development—I conducted Time Series analysis to identify trends associated with alerts across five threat levels, resulting in a remarkable 73% reduction in false positives. Furthermore, I utilized K-Means Clustering and DBScan to analyze and identify trends for these threat levels, ultimately establishing an optimal baseline for our models. Notably, I also contributed to the development of an ETL pipeline and a Spark Streaming application responsible for managing data from 500+ tenants using Databricks and Pyspark. Additionally, I constructed a Databricks Job that runs on 4000+ tenants hourly, leveraging MultiThreading to enhance efficiency by 59%. Your impactful work demonstrates a commitment to innovation and sustainability within the technology landscape.

As an Undergraduate Research Assistant under the guidance of Prof. Sihang Liu, I am actively engaged in a pioneering project aimed at addressing the critical challenge of reducing carbon emissions from cloud computing centers. My work involves the development and meticulous fine-tuning of neural models, harnessing their power to predict carbon emissions and energy outputs within complex cloud environments. To ensure the precision and resilience of our predictive models, I've conducted thorough data analysis and pre-processing, gathering vital information from cloud infrastructure, energy grids, and environmental sources. This holistic approach not only drives innovation but also contributes to our collective efforts in promoting sustainability within the technology landscape.

During my time working at SmartGene, a forward-thinking startup, I had the incredible opportunity to contribute to a project that aimed to democratize education worldwide. SmartGene's platform offered students from diverse backgrounds access to world-class tutoring and assistance at absolutely no cost. As a key team member, I took on the responsibility of building their website, ensuring it was user-friendly and efficient in connecting students with tutors. Additionally, I actively participated in shaping their business venture plans, which revolved around creating a sustainable model for providing quality education to all. It was a fulfilling experience to be part of a team dedicated to making a positive impact on global education.

In one of my most impactful projects, I harnessed the power of Azure Language Services and Azure Bot Services to develop a highly reliable chatbot tailored for emergency room (ER) patients. The primary goal was to provide patients with timely and trustworthy information during critical moments. I successfully integrated Azure QnA maker, accurately categorizing over 300 common diseases and conditions based on their symptoms, spanning from severe cases to milder conditions. This ensured that patients received the most relevant information and assistance when they needed it most.

To further enhance the bot's capabilities, I leveraged the OpenAI API, utilizing modern-day Language Model Models (LLMs) and conducting a comparative analysis with Azure's model. Through meticulous Prompt Engineering and Chain of Thought Reasoning, I achieved an impressive 95% accuracy rate in categorizing, diagnosing, and resolving over 500 medical cases. This project not only exemplified my technical expertise but also demonstrated my commitment to leveraging cutting-edge technology to improve healthcare accessibility and patient outcomes. It was a rewarding experience to contribute to the development of a solution that could potentially save lives in emergency medical situations.

As a Software Developer at the University of Waterloo Satellite Design Team, I took the helm in creating two innovative web applications that played a pivotal role in satellite communication. Leveraging my expertise in React and Firebase, I led the development of these applications. One notable accomplishment was the design and implementation of an ARO Requests page, a central hub for all incoming and outgoing queries related to the satellite. This strategic move resulted in a remarkable 40% reduction in response time, enhancing the efficiency of our satellite operations. Furthermore, I engineered a secure login form using React and Firebase, which not only bolstered security but also streamlined the transmission of satellite requests, leading to a 20% improvement in overall efficiency. These experiences were not just about coding; they were about driving technological advancements in the field of satellite design and communication. I am proud to have played a pivotal role in advancing the team's capabilities and contributing to the success of our satellite missions.