Engineer, full-stack developer, creator. 20+ years writing code across the full stack, from product interfaces down to the systems layer where most people stop looking.

AI is part of my development workflow for research, learning, documentation, and code generation where appropriate. All architectural decisions, validation, testing, and final implementations are my responsibility. I treat AI as an assistant, not an authority.

Encoding Station Client-side multi-transform encoder/decoder for CTF work and security research. Parallel mode runs all transforms simultaneously. Pipeline mode chains them in drag-reorderable steps. Base64, Hex, URL, HTML entities, ROT13/47, Binary, XOR — all processed locally, no server, no build step. JavaScript · React · CTF · Security Research

pretty-good-encryption Client-side encryption tooling built on the Web Crypto API. JavaScript · Web Crypto API · Cryptography · Security Research

Glasswally Real-time LLM distillation attack detection via eBPF kernel uprobes. Hooks into inference processes at the kernel boundary. Fingerprints query patterns. Alerts before a model theft completes. MITRE ATT&CK mapped. Rust · eBPF · Detection Engineering · AI Security

ghostwire Passive USB device fingerprinting and anomaly detection daemon. Detects BadUSB implants, HID injection devices, and hardware anomalies at the enumeration layer — before the OS finishes booting the driver. Silent. Rust. No kernel module required. Rust · Linux · Endpoint Security · Hardware Fingerprinting

Milly Local LLM with a built-in security layer. HMAC-signed memory, TF-IDF RAG, Guardian injection detection, and a tamper-evident audit trail. Zero egress by design. 157 passing tests covering memory integrity, injection resistance, and audit logging. Python · Ollama · AI Security

passive-intel Passive threat-intel console for defenders. CISA KEV, URLhaus, and Feodo Tracker in one UI — actively exploited CVEs, live malware URLs, and botnet C2 IPs. No API keys. No cost. Includes a blocklist endpoint you can feed directly into nginx, ufw, or Cloudflare. JavaScript · OSINT · Threat Intel

EverythingOS Security-first multi-agent framework for autonomous systems. ModelGuard input/output validation. DecisionLedger audit trail. NIST AI RMF aligned. 81/81 tests passing, zero TypeScript errors. Built under the Robots For Peace framework. TypeScript · Autonomous Agents · AI Governance

Rust · Python · TypeScript · JavaScript · React · Three.js · eBPF · mitmproxy · Ollama

Kernel-level tooling · Detection engineering · AI security · Autonomous systems · Creative coding

The purpose is not just to build tools — but to fully understand how they work and why. Running them is not enough.