Engineer execution.

Selected projects.

Technical implementation of Opaque Clarity security — from bare-metal infrastructure to autonomous agents.

Infrastructure

Project GhostNode

A segmented home-lab infrastructure designed to simulate enterprise attack surfaces and test failure-aware architectures under realistic conditions. GhostNode is the proving ground for Opaque Clarity doctrine: every detection, every logging decision, and every network boundary gets stress-tested here before it goes anywhere near a production recommendation.

Focus areas: logging strategy and evidence path integrity, network segmentation and lateral movement detection, local AI integration for behavioral baselining, and secure storage for key material and forensic artifacts.

Proxmox Ollama (Local AI) Splunk Network Segmentation Secure Storage
Development

WeanTech Swarm

A distributed, edge-based multi-agent research system exploring adaptive cyber operations under concept drift and adversarial conditions. Agents use online learning and limited peer communication to characterize evolving network behavior — supporting both defensive monitoring and controlled red-team emulation without reliance on continuous centralized analysis.

Autonomy is safety-bounded and auditable. No agent acts outside its defined envelope without escalation.

Python Multi-Agent Systems Online / Continual Learning Adversarial Evaluation
AI Research

OctopusONN

A biologically-inspired neural architecture modeled on the octopus nervous system. Semi-autonomous arm modules using Closed-form Continuous-time (CfC) cells coordinate via event-driven signaling rather than centralized control — designed to evolve toward a federated, emergently coordinating system for adaptive security applications.

Current focus: architectural stability under concept drift and controlled emergent behavior in adversarial environments.

CfC Cells Federated Learning Biologically-Inspired AI Event-Driven Coordination
AI / ML

Applied ML: Classification & Detection

Applied computer vision (CIFAR-10) and phishing detection models — moving from theory to practical model training, evaluation, and deployment. Emphasis on adversarial robustness and behavioral signal extraction over surface-level pattern matching.

PyTorch TensorFlow Jupyter Classification Adversarial Robustness
Architecture

Opaque Clarity Security

The WeanTech doctrine applied end-to-end: systems designed to be opaque to adversaries (high friction, minimal signal leakage) and transparent to defenders (full observability, clean evidence paths). Architecture is threat-modeled first, not hardened after the fact.

Security Architecture Threat Modeling SOP Design Incident Response
Live tool

ThreatLumen

A live threat intelligence dashboard built for daily situational awareness — not a demo, not a portfolio piece. This is what I actually use to monitor threat feeds, track indicators, and maintain situational clarity on the current threat landscape.