People / Students / Class of 2026
Keyu Yan graduated from Rutgers University with a major in Mathematics and a minor in Computer Science. Throughout his academic journey, he developed a rigorous analytical foundation and gained proficiency in Python, R, and SQL, along with modern machine learning libraries. This training equipped him to translate ideas into practical, working systems. In industry, at HFT Investment Management, he applied K-Means clustering to transaction streams, cutting manual review time by 95% and uncovering 30+ non-compliant patterns. Building on this, he redesigned credit-risk evaluation with logistic regression and decision trees, reducing processing time by 90% and improving model accuracy metrics significantly. He later enhanced trading signal forecasting through random forests and gradient boosting, providing more reliable decision support for portfolio management. Complementing these efforts, his research focused on scale and reliability: engineering TB-level datasets, applying PCA and mutual-information screening for feature selection, and developing a deep learning–XGBoost pipeline capable of processing 10M+ rows, which improved predictive accuracy by about 20%. He hopes to use data science to drive innovation, improve decision-making, and tackle real-world challenges through advanced data analysis and machine learning. He views the MLDS program as the ideal next step—offering modern machine learning training, hands-on work with real data, and guidance from experienced faculty, along with a diverse cohort to broaden his perspective. With this foundation, he looks forward to refining his technical expertise, gaining practical insights, and moving closer to his ambition of becoming a data scientist.
