Dawsonville, Georgia, United States
I’m a software engineer focused on building practical, human-centered web applications with React and modern JavaScript. Right now I’m developing education-focused tools, including a learning app concept that tracks growth, highlights mistakes as learning signals, and supports students without replacing their thinking. I’m especially drawn to problems where UX, clarity, and motivation matter — turning complex ideas into simple interfaces, improving learning flows, and designing features that guide users instead of overwhelming them. One example of impact: I’ve been building and refining a student-focused learning platform with game-style progression, streak systems, and feedback signals that reward effort and improvement — aimed at helping academically struggling learners stay engaged and find their own voice. I like building tech that teaches, supports, and empowers.
Performed high-level technical auditing and adversarial testing of frontier Large Language Models (LLMs) and agentic coding tools. Focused on evaluating the reasoning capabilities, architectural integrity, and execution accuracy of AI models when navigating complex, production-grade codebases. Core Responsibilities & Achievements: Agentic Workflow Auditing: Evaluated the performance of terminal-based coding agents (Gemini CLI, Claude Code) in managing multi-step software engineering tasks, including file system manipulation, environment configuration, and version control. Codebase Stress Testing: Designed sophisticated "trap" prompts to test model robustness against underspecified requirements and technical ambiguity. Utilized high-consequence libraries like Zustand and Marshmallow to identify failures in state management logic and architectural integration. Failure Mode Analysis: Documented and categorized critical model errors, including Verification Failures, False Claims of Success, and Overengineering. Provided root-cause analysis for instances where AI-generated solutions deviated from established design patterns or introduced unnecessary complexity. Environment & Containerization Management: Orchestrated localized development environments using Docker, WSL, and Linux (Ubuntu) to validate model-generated code. Resolved complex dependency conflicts, such as Node.js ESM/CommonJS package mismatches, ensuring environment parity between local and containerized states. Execution Flow Tracing: Conducted deep-dive audits of execution flows and state-mutation patterns to detect subtle hallucinations in AI-generated documentation and logic explanations. Technical Stack Integration: Leveraged a technical background in React, Python, and JavaScript to assess model proficiency in modern frontend frameworks and backend automation scripts.