Ryan Holmes

I’m Ryan Holmes, a machine learning researcher with a track record of developing innovative AI solutions for complex, high-dimensional, and time-sensitive data. My expertise spans the full ML pipeline—data preprocessing, feature engineering, model development, evaluation, and deployment—with a strong focus on interpretable, updateable, and production-ready systems.

Currently, I work as a Machine Learning Research Scientist at Ask2.ai, where I design algorithms and frameworks that extend classical methods like PCA, K-Means, and clustering into robust, time-aware systems capable of adapting to streaming data without sacrificing stability. My work includes building multi-modal LLM/RAG pipelines, advanced document classification frameworks, and custom visualization tools that transform high-dimensional embeddings into actionable insights using t-SNE, UMAP, and interactive Python animations. I lead cross-functional research projects that blend rigorous statistical modeling with large-scale AI architectures, always prioritizing explainability, scalability, and measurable impact.

I excel at transforming messy, real-world datasets into reliable models that deliver meaningful results—combining technical rigor with clear, intuitive outputs. I believe data visualization is one of the most important bridges between human understanding and machine intelligence, making complex models transparent, explainable, and ultimately more trustworthy. The animation below is one example of how I use visualization to reveal model behavior and uncover patterns that raw metrics alone often miss.

I earned my Master of Science in Financial Engineering from Columbia University, where I developed a strong foundation in advanced quantitative methods, machine learning, and computational finance. The program challenged me to integrate rigorous mathematical modeling with real-world financial applications. Through courses such as Deep Learning, Reinforcement Learning, Monte Carlo Methods, and Algorithmic Trading, I gained hands-on experience designing and implementing models that operate under uncertainty and time dynamics. My technical skill set was further refined through Programming for Financial Engineering (C++) and Optimization, equipping me to build robust, high-performance systems in both research and production settings.

Prior to Columbia, I completed my undergraduate studies at Butler University, earning a Bachelor of Science in Finance with minors in Mathematics, Actuarial Science, and Art + Design. This multidisciplinary background gave me a well-rounded perspective on both the quantitative and creative aspects of problem-solving. I pursued advanced coursework in Real Analysis, Financial Mathematics, Econometrics, and Financial Derivatives, laying a solid analytical and theoretical foundation.

Outside of my research, I enjoy exploring a variety of creative and strategic pursuits. I work in both acrylic and oil painting, and I’m an avid player of strategy and tabletop games such as Warhammer and Magic: The Gathering. I also have a passion for history, often diving into documentaries that bring the past to life, and I find joy in hands-on creative outlets like building with LEGO. These hobbies give me a balance between analytical thinking and artistic expression, fueling both my problem-solving and creativity.