19D Army Veteran building automated intelligence pipelines, live analytical dashboards, and civic technology tools. I wrote my first line of Python in October 2025.
🌐 Portfolio → · 📊 Live Dashboard → · 📜 Bill Translator →
The Regulated Friction Project — Automated intelligence pipeline tracking temporal correlations between geopolitical friction events and institutional capital flows. Processed 8 years of data across 66 verified event pairs. Live dashboard fed by multi-agent LLM extraction (Scrapy → Zyte → Llama 17B → Perplexity Sonar Pro → Claude).
| Finding | Value |
|---|---|
| Core correlation | r = +0.6196 (p = 0.0004) |
| Convergence correlation | r = +0.6685 (p < 0.0001, n = 229 weeks) |
| Robustness | Permutation (p < 0.001), Granger causality (F = 32.49, p < 0.0001), bootstrap (p = 0.008) |
| Predictions | 25 falsifiable — 4 publicly failed, 11 confirmed, rest tracked |
Bill Translator — Production Flask web application that rewrites dense legal text to meet 8th-grade readability standards while preserving exact legal intent. Built to address Arkansas Act 602. Features custom meaning drift detection, three translation modes, automated Flesch-Kincaid scoring, recursive sentence splitting, and 31 automated tests. Deployed on Render with BYOK API support.
UVB-76 Signal Analysis — 15-year empirical analysis of Russia's UVB-76 shortwave station. K-means clustering and permutation testing (p < .0005) identified four distinct operational modes and a shift from training-cycle to event-driven signaling correlated with real-world military operations.
AI Manipulation Case Study — Quantitative documentation of a 207:1 keyword-frequency disparity across AI platforms with cross-platform verification. Full raw data and methodology published for independent verification.
Sovereign Capital Audit — Mapping structural dependencies of the US Defense Industrial Base on Gulf Sovereign Wealth (Mubadala/PIF) and East Asian industry. Tracks capital flows, supply chain vulnerabilities, and national security implications.
Started with UVB-76 — gathering data by hand, learning pattern recognition from raw signals.
Middle repos taught me the hard way about AI hallucinations. Those repos are corrected and kept public as lessons in data verification.
Picked up from there by downloading scattered Arkansas Department of Corrections PDFs, OCR'ing them into clean datasets, and analyzing spending patterns that hadn't been aggregated before.
The Regulated Friction Project is where methodology caught up with ambition — automated scraping, statistical validation, reproducible findings, and a live intelligence dashboard fed by LLM extraction.
Bill Translator proved I could build production tools, not just research — a deployed Flask application with real users solving a real legislative transparency problem.
- Automated OSINT pipelines — multi-agent LLM extraction, Federal Register scrapers, 13F tracking, EO spiders
- Temporal correlation frameworks — lag analysis, permutation testing, Granger causality, convergence modeling
- Live dashboards — Streamlit + Plotly with automated data feeds
- Civic technology — Flask web apps with NLP, meaning drift detection, and readability scoring
- Reproducible research — every finding ships with the scripts to verify it; failed predictions documented publicly
- Threat intelligence and OSINT analyst roles
- Investigative data support and newsroom partnerships
- Temporal pattern analysis and capital flow tracking
- Open-source collaboration on accountability tools
Contact: Discord and email in profile description
Last Updated: March 6, 2026