Kwame Osei-Tutu
University of Massachusetts Amherst
Mathematics and Statistics
Biography
I'm a 4th-year Math PhD at UMass Amherst optimizing ML models via mixed precision techniques. I'm currently working improving parameter estimation and prediction in Gaussian processes, with plans to extend these optimizations to other ML architectures. My focus is on leveraging HPC clusters to accelerate model training/inference while maintaining accuracy—bridging numerical analysis and scalable ML systems.
Academic Status
PhD Student - 4th
Research Area/Department
Applied Mathematics; Data Science; Machine Learning/AI; Mathematics
Major/Specialty
Machine learning, gaussian processes, high-performance computing
Degrees Earned or in Progress
Earned a Bachelors of Arts in Mathematics and Statistics Currently a 4th year Math PhD student
Academic Preparation
I have taken a bootcamp in Data Science with the Erdos Institute during the summer. My PhD program is in applied mathematics so I have a wide range of experience in quantitative topics. In terms of computer science experience, I recently took a course in Optimization and currently completing a machine learning course where I'm learning the concepts that are pervasive in modern machine learning.
Research/Academic Interests
Optimizing ML models via mixed precision techniques. Currently improving parameter estimation and prediction in Gaussian processes, with plans to extend these optimizations to other ML architectures. Focus on leveraging HPC clusters to accelerate model training/inference while maintaining accuracy—bridging numerical analysis and scalable ML systems. By the time the summer internship starts, I probably would have published at least one paper on this topic.
Computational and Data Science Areas
Applied Computer Science; Applied Mathematics; Artificial Intelligence and Intelligent Systems; Computer Science; Performance Evaluation and Benchmarking; Statistics and Probability
Motivation
From a young age, I've been fascinated with both mathematics and computers—mathematics because I believe it's the language of the universe, and computers because of their ability to create solutions that previous generations would have marveled at. It is that passion that motivates me everyday in my research to read and attempt to push the boundaries of knowledge and tools at humanity's disposal. My interest in the cutting edge of technology is what drives me to want to work with research labs that are, even in the smallest of ways, pushing the envelope in any field at the intersection of math and computing. Therefore, the summer 2026 projects that are focused on AI and HPC are perfectly aligned with my interests and that's why I've felt compelled to apply.
Lightning Talk Title
Accelerating Machine Learning with Mixed Precision
Keywords (Maximum 20 words)
Gaussian processes; mixed precision; parallel computing; python; R