I am Hunter, an AI Research Scientist at Roots.ai based in Jersey City, NJ. I build systems that bridge the gap between academic rigor and production-scale machine learning. With over 16 years of software engineering experience, my work focuses on applying robust representation learning to complex, unstructured domains. This ranges from extracting structured data from messy business documents to forecasting chaotic physical systems.
My journey in software began with indie game development, where I built physics-based engines and cross-platform experiences. During my undergraduate studies at Drexel University, I transitioned into machine learning and deep learning.
Working at SAP’s Conversational AI Labs during the release of the original GPT model sparked a deep fascination with computational linguistics and document processing. I spent my early career exploring the social impacts of generative text and conversational dynamics.
I took this foundation in generative modeling to Harvard University for my PhD studies. I applied architectures originally designed for language, like transformers and VAEs, to scientific simulation and molecular dynamics. This work brought a new level of mathematical rigor to my engineering. I successfully passed my qualifying exams and advanced to PhD candidacy. During this time, I realized my core values align most strongly with open-source development and immediate real-world impact. I chose to leave with my Master’s degree to return to the applied NLP space.
Today, at Roots.ai, I lead initiatives in parameter-efficient fine-tuning and distributed training for multimodal datasets. Recently, I spearheaded the training and release of GutenOCR, a terabyte-scale open-source Vision-Language Model project featuring open weights, open code, and open data. My goal is to build systems that automate tedious workflows and create tangible value for people, processing millions of documents with high accuracy and reliability.
What You’ll Find Here
This site showcases my research and shares practical ML knowledge. You’ll find:
- Research: Publications on LLMs, document processing, scientific ML, and AI safety (including COLING 2025, EMNLP, AIES)
- Posts: Technical tutorials on foundation models, generative AI, and scientific computing
- Projects: Open-source implementations and experiments, including my early game development work
- Notes: Reference material on datasets, methods, and tools
Let’s Connect
I am always open to discussing new ideas, sharing insights on scaling ML systems, and exploring the intersection of AI and the physical sciences. Whether you are building production pipelines, conducting research, or just want to talk shop about VLMs, open-source datasets, or indie game dev, feel free to reach out via email or connect with me on LinkedIn.