I've taught the last 10 years as a tenured Biology Professor and have 10 years in pharmacology research with increasing amounts of coding over the last 20 years. I'm ready to switch roles that focus on custom R&D for hard problems. Deeply curious, love to learn. I'm eager to meld my experiences in research, pharmacology, custom algorithm design, system architecture and data analysis pipelines to solve some HARD problems.
This is me:
I use LLM pair-programming to accelerate design, pressure-test architectural designs, and achieve extreme development velocity
Website: gitgalaxy.io | PyPI: pypi.org/project/gitgalaxy/
Description: An AST-free, LLM-free heuristic knowledge graph engine for deep repository intelligence. A custom engine forked from the BLAST (DNA scanning) algorithm. Map, understand, secure, and modernize enterprise codebases across 50+ languages at extreme velocity.
Key Implementations:
- Full Architectural Teardowns: The engine parses coding intent across 50+ languages. Translates structural regex hits into calculated risk exposures to map information flow, system design, network analyses and architectural drift. (Examples: Apollo 11 source code, IBM CICS benchmarks).
- Zero-Trust Security & Threat Inference: The engine focus is tweaked to now isolate logic bombs, memory corruption, and hidden vulnerabilities. It utilizes C-backed Shannon Entropy math to catch obfuscated malware and an embedded XGBoost model to classify specific payloads like Trojans and Droppers.
- Supply Chain Defense: High-velocity perimeter scanner that monitors every file before download. Actively searches for hidden payloads, magic byte mismatches, and extreme entropy in compiled binaries.
- Legacy Modernization: Automated, deterministic pipelines that map, slice, and refactor enterprise architectures, including COBOL to Java and COBOL to COBOL.
Status: Functional & Deployed with website, PyPI (14k+ downloads), and git community usage.
Website: helpingfarmersfarm.com
Description: A platform built to connect eager community members with local agriculture for work-share CSA programs. It gives farm managers the tools to seamlessly oversee volunteer rosters, track seasonal work commitments, and manage digital liability waivers without the hassle of spreadsheets.
Status: Functional & Deployed - 1 farm using system for 3 years.
Description: A lightweight, distributed SCADA middleware for physical automation over RS-485. Features an asynchronous Python Digital Twin host (RP5) controlling deterministic, bare-metal MicroPython nodes (RP2350). Currently implemented as a small parts sorting system.
Website: fastmathfacts.io
Description: A gamified math practice engine designed to reward both mastery and gritty determination. This uses a dual scoring system that mathematically separates mastered repetition from hard work. It features dynamic game modes to target weak spots, millisecond timing with instant feedback and different analytic boards to prove the student is growing with practice.
Status: Functional & Deployed with website. People practice and submit to my absurd international leaderboard.
Website: github.com/squid-protocol/sorting_evolution_algorithm
Description: A custom-built Genetic Algorithm (GA) and Machine Learning pipeline designed to procedurally generate, simulate, and optimize physical vibrating sorting mechanisms for small parts. Because standard optimization algorithms fail when applied to chaotic 2D rigid-body physics, this engine utilizes a three-pillar architecture: a multi-threaded Web Worker pool running headless matter.js physics simulations, a Python/Dash dimensionality reduction pipeline (scikit-learn PCA) that maps AI strategies into a 3D fitness landscape, and a rigorous Monte Carlo stress-tester to eliminate statistical flukes.
Status: Functional. The project successfully evolved optimum design specifications, which are currently implemented in the physical SCADA (meow turtle) hardware project above.
I love building and enjoying working with others. If you've got a project, I'm happy to hear about it.