Selected engineering work.
I'm an automation developer, Navy veteran and CompTIA Security+ holder who designs and ships systems end to end — from an algorithmic trading platform to deployed SaaS products, AI automations, and a home security lab. This page is the technical detail behind that work, for recruiters and engineers; the Work page next door is written for customers instead.
Open to: IT support, security, automation and software roles — remote, or on-site and hybrid in Hampton Roads, VA.
CompTIA Security+ ce
Issued by CompTIA · Verify on Credly →
AMP Titans Capital
Algorithmic trading platform — A production trading system that gives automated strategies queryable memory of their own history.
The problem
Automated trading bots tend to make the same categorical mistakes on repeat, because every decision is made in isolation — nothing in the system remembers whether a similar setup worked before.
What I built
A multi-service platform: a FastAPI backend and a risk-managed execution engine trading live through the Ironbeam broker API, a Next.js dashboard for monitoring bots in real time, and — the core of it — a Neo4j knowledge graph built with Graphiti that extracts entities from narrative trade logs, so a strategy can query "has this bot and condition combination worked before?" before it acts.
Proprietary system — architecture and stack shown here; strategy logic and code are not public.
Built, deployed, and running.
The Cleaners Place
The all-in-one operations platform house-cleaning businesses run their day on.
Visit live →Full case studies — problem, approach, and results — on the Work page.
Applied AI, tooling, and side projects.
Lead Intent Monitor
Finds public conversations where prospects are already talking about a problem your product solves, scores them, and surfaces the ones worth responding to today.
Built first for The Cleaners Place — a FastAPI backend with an AI classifier scores conversations for intent and relevance, sitting behind a password-gated admin dashboard. The scoring engine (keywords, sources, classifier) is generic and reconfigurable for other businesses.
ForgeDeck
A Windows-first spatial command center for orchestrating multiple coding agents — Claude Code, Codex, terminals, browsers — from one persistent workspace.
An infinite pan/zoom canvas (React Flow) where each agent, terminal, and browser session is a draggable node instead of a buried tab. Built milestone-by-milestone against a written spec; currently at the working-shell stage.
RideSight
A computer-vision coach for bull riding — upload a ride video, get evidence-grounded feedback on body position and technique.
A custom fine-tuned rider/bull detector feeds a SAM 2.1 tracking pipeline and 2D pose estimation, run frame-by-frame through a local GPU inference server. Uncertain detections are withheld from riders and routed to a human review queue instead of guessed at.
Systems that do the busywork.
NoteBot
A Discord bot that records team calls and turns them into finished documentation without anyone touching it.
Captures the call, transcribes it with Whisper (GPU, with CPU fallback), summarizes it with an AI model, writes a Google Doc, and logs the action items as Airtable tasks. Includes a FastAPI dashboard and a Windows tray app, and handles recordings of about three hours.
Airtable and workflow automation
Operational systems built on Airtable and wired together with Make and n8n.
Patient trackers, financial reporting for medical facilities, general financial tracking, lead tracking, and an automated blogging pipeline. Flows run through Make and n8n, with Airtable as the database and interface layer.
Home security lab.
Hands-on practice behind the Security+: building the tools a defender uses and watching how attackers behave.
Honeypot lab with Wazuh SIEM
A deliberately exposed Linux VM wired to a Wazuh SIEM, so attacker activity is captured and kept off the machine being attacked.
Built with Vagrant and VirtualBox. Activity is forwarded to Wazuh, so the evidence survives even if the honeypot itself is wrecked. That's the collect, detect and preserve loop a SOC analyst works in.
MalVault malware-analysis lab
An encrypted malware-analysis lab and sample vault for studying how malware behaves.
Samples are stored encrypted at rest, using AES-256 with PBKDF2 key derivation, all scripted in PowerShell.
Public repos, open to browse.
EHR platform →
A multi-tenant ambulatory EHR, practice-management and revenue-cycle platform, built from a written spec. All data is synthetic.
FounderOS →
An AI-assisted back office for founders: contracts, compliance, tasks, customers and money, each tool a separate module.
RentNest →
Property management for small landlords: leases with e-signatures, rent payments, maintenance and a tenant portal. No longer hosted; the code is here to read.
The Cleaners Place for Claude and ChatGPT →
The public plugin and workflows that let cleaning-business owners run their business by chatting with an AI assistant.
Proxy Collector & Validator →
Standalone Python app that collects, validates, and exports proxies — Tkinter GUI over a Flask backend.
CursorGPT →
Select any region of your screen, OCR it, and get a concise GPT answer in a floating tooltip — for quick lookups and code explanations.
Building something interesting?
I'd like to hear about it — reach out, or look through the rest of the code on GitHub.