AI Developer, building trust and verification infrastructure for autonomous agents
I build small, sharp tools for problems I actually hit. Most of my current work sits near one question: what happens when an agent is asked something the internet doesn't know?
I finished an M.S. in Artificial Intelligence at Worcester Polytechnic Institute. My thesis, CHIPNet, built deep neural networks for precisely locating corners on semiconductor chips, a three-stage hierarchical architecture advised by Prof. Ziming Zhang and developed in collaboration with Teradyne.
I now work as an AI Developer on production LLM systems. Outside of that I ship open-source tools and spend most of my thinking on where agent infrastructure is heading, particularly the parts concerned with trust.
The problem I keep coming back to. The public web isn't the world's knowledge: it's the published subset. Enormous amounts of true, valuable, verifiable knowledge are never written down anywhere. A search agent can't retrieve any of it, and it usually doesn't say so. It answers anyway.
Accessibility review for the roughly 60% of WCAG criteria automated tools can't check, the ones that need a judgment call about meaning rather than a rule about markup.
Checks how crowded an idea already is on GitHub before you spend a weekend rebuilding something that exists. Turns "did anyone do this already?" into a five-second habit.
GitHub trending search that correctly handles the Search API's undocumented behaviour, including the one where grouping parentheses return zero results with no error at all.
Context understanding and parameter-efficient fine-tuning for generative models.
IEEE ICICIT 2024 · 6th International Conference on Inventive Computation and Information Technologies · April 2024
A matching system connecting students with internship opportunities.
ICDSMLA 2023 · International Conference on Data Science, Machine Learning & Applications · December 2023
A three-stage hierarchical deep neural network for precisely locating corners in semiconductor chip images, reaching 99.06% patch detection accuracy. Developed in collaboration with Teradyne.
M.S. Thesis · Worcester Polytechnic Institute · Advisor: Prof. Ziming Zhang · 2025
Production LLM systems: building AI agents and workflows, and researching generative AI methods for enterprise applications.
2025–present
Led the university-industry research project behind CHIPNet under Prof. Ziming Zhang: architected a three-stage hierarchical CNN in PyTorch for locating corners on semiconductor chips, reaching 93.4% detection accuracy, a 35% improvement over traditional computer-vision methods. Built the full pipeline end to end, including data augmentation and a real-time inference system.
2025
Computational engineering and real-time embedded systems.
2025
Thesis: CHIPNet, deep neural networks for locating corners on semiconductor chips, advised by Prof. Ziming Zhang.
2024–2026
Built and maintained the university festival platform for two years; co-authored a research paper on large language models as an undergraduate.
2020–2024
Happy to talk about agent infrastructure, retrieval systems, or anything at the seam between research and production. The fastest way to reach me is email.