Adaptive resonance
Fuzzy ARTMAP learns new categories without forgetting old ones. We study how automated tuning and the decision readout change its accuracy when every compared method gets the same compute budget.
Independent research
Adaptive and explainable machine learning, tested the hard way: equal compute budgets, plans fixed in advance, and every result reported, including the nulls.
Focus
Fuzzy ARTMAP learns new categories without forgetting old ones. We study how automated tuning and the decision readout change its accuracy when every compared method gets the same compute budget.
Classifiers whose decisions can be read and audited by the people accountable for them, especially in regulated settings such as health and security.
Fair comparisons are the hard part. We fix the analysis plan before the main run, log deviations, and publish code and full results.
Papers
An interpretable Fuzzy ARTMAP approach to document review.
CEUR Workshop Proceedings, vol. 3832.
Sole-authored and not published; reviewed by my doctoral chair. Studies automated tuning and decision readout of Fuzzy ARTMAP on 253,680 CDC survey records. PhD conferred July 2026.
How we work
About
Researcher and technical leader at the intersection of infrastructure, security and AI. PhD in Computer Science (AI/ML) from National University, conferred July 2026.
More than twenty years of building and running enterprise networks: data center, SD-WAN, and hybrid cloud for HIPAA-regulated workloads. Managed engineering teams and ran an independent consulting practice before turning to research.