Technical documents
Both documents below were produced for a client engagement under NDA. They are described accurately here and are available on request; a public version would need a redaction pass and sponsor sign-off first.
Knights Counsel: Technical White Paper, v1.0
Co-author · sole author, system design and behavior specification sections · August 2026
A full technical account of the system's architecture, behavior specification, evaluation methodology, and results, prepared for enterprise IT and the client. The sections I wrote cover the two-agent orchestration topology, the thirteen numbered behavioral requirements, the grounding and citation enforcement model, and the configuration baseline treated as version-controlled.
Knights Counsel: Final Design Document, v2
Contributing author · sole author, system design section · 2026
Requirements, architecture, and design rationale for the sponsored capstone. Includes the behavior test set and the architecture decision record covering the six-agent to two-agent reversal — written up as a record rather than deleted, so the reasoning survives for whoever picks the project up next.
Architecture decision records
Author · ongoing through the engagement
A running set covering platform selection, delivery surface, access model, session memory, citation rendering, prompt-language requirements, naming enforcement, and fallback strategy. Each captures context, decision, consequences, and status — including the records that document reversals of the team's own earlier decisions.
Research work
Hardening Build Systems
Undergraduate research assistant · University of Central Florida · Oct 2025 – May 2026
Empirical research on the security and reliability of open-source software build systems. I diagnosed build failures by hand across six build systems, and built the tooling that converts raw experimental output into LaTeX-ready tables and pgfplots figures used in project reporting. The failure categories that came out of that work are what the Build-Recipe project is now built to measure.
Human Factors Group Study
Research assistant · University of Central Florida · Feb 2025 – May 2026
A study of how programmers debug code collaboratively in groups. I ran experimental sessions and instrumented participants with biometric equipment measuring heart rate, respiration, skin conductance, gaze, and prefrontal cortex blood oxygenation. Worked under human-subjects research protocol with the associated ethics training completed.
Research interests
Applied AI systems and the evaluation methods that validate them: LLM agent design, retrieval grounding and citation enforcement, guardrail and refusal behavior, and the measurement problem of separating genuine capability from fit to a chosen benchmark.
Secondary interests in empirical software engineering — particularly build-system reliability and reproducibility — and in human-computer interaction at the level of accessible, well-structured interfaces.