25+ years in tech — from early networking to containerized microservices, IT operations at scale, automated trading systems, and AI-driven tooling. I build systems that run without excuses, and every project here is deployed and running.
Started in tech through high-school ROP programs and never looked back. 25+ years later I've worked through bare metal, the cloud-migration era, and now build containerized, event-driven systems that run in production every day.
My focus sits where IT operations meets building things — internal tools, support and reporting dashboards, e-commerce pipelines, automated trading systems, and the AI-driven automation that ties them together. For the past three years I've worked AI-natively, using modern tooling to ship faster and scale support without scaling headcount.
I believe in shipping production code, not slide decks. Every project below is deployed and running.
Production systems handling real traffic, real money, and real users — grouped by what they do.
Full-stack operations hub for One N Only Entertainment. Real-time WooCommerce sync, cash-flow forecasting, payment-plan management, a vendor RBAC portal, and Meta Ads automation — all in a single containerized deployment.
A suite of custom WordPress/WooCommerce plugins: a bundle-receipt fix, payment-plan checkout with Stripe invoicing, and a real-time order-management module wired into the dashboard.
Festival social and roommate-matching app. A multi-container Docker stack — API, frontend, Redis, and Postgres — secured behind Cloudflare Zero Trust with QR-code entry.
Community photo-sharing app for a Las Vegas festival weekend. Cloudflare Zero Trust auth, a diamond-grid UI, real-time uploads, and a staged rollout plan — integrated into the main dashboard.
Six small automated bots on a decentralized perpetuals venue, one per wallet and timeframe, wrapped by independent services: a liquidation-aware stop guardian, a paced market-data proxy, a position monitor, and a fills relay. Signals arrive from TradingView through Cloudflare tunnels and a fan-out; every fill is recorded to a public 60-day ledger, losses included. A hobby engineering project, not a fund.
A falsification-first research stack: every strategy idea is pre-registered, replayed on six years of data with a random-entry placebo and a drift-preserving bootstrap null, a positive control through the same pipeline, and a held-out window. ~53 verdicts in a public-style ledger so nothing gets re-chased; a market feature library that measures what state predicts (move size: yes; direction: no). "Nothing here" is an allowed answer, and usually the honest one.
The ops dashboard behind Cloudflare Access: live positions, take-home P&L after fees, bot health, risk guardrails, alert-pipe status, funding and structure radars, a Stripe card — regenerated every two minutes from a snapshot the feed already holds. Around it: an 82-check morning audit to Telegram, a patch log that is derived from live container images and refuses to publish leaky copy, and a 1-second order-flow recorder.
A second, isolated box that scans five majors across seven timeframes every hour with structure detectors (break of structure, change of character, fair-value gaps, order blocks, liquidity sweeps) and publishes a dashboard plus a per-timeframe bias summary that the command center pulls over a private VPC. Built from a zip to a live tunnel in one session, four shipped bugs fixed on the wire.
An AI co-pilot that runs operations from a persistent, git-tracked memory of every deploy, incident and research verdict (300+ entries since April 2026); multi-agent workflows with adversarial review stages for audits, designs and documents; a retrieval "brain" over the project's own refutations that surfaces "we already killed this" before anything else; and an advisor whose numbers are audited by code that fails closed. It holds no keys and cannot trade — the human decides what goes live.
A Pine Script trading indicator with a configurable status-box display and a structured JSON webhook payload architecture that streams real-time alerts into a dedicated server for tracking.
A multi-agent market-observation system for crypto perpetuals. TimescaleDB and FastAPI, with a structure-detection agent that uses fractal pivot logic to read market structure in real time.
Methodology-first trading research. A custom entry-skill testing framework using Bonferroni-corrected permutation testing for honest signal validation — publishing null results as readily as wins.
The original automated trading engine on Solana via Drift Protocol — a 15-module Python engine with webhook routing, SaaS tiers, Phantom wallet auth, and an investor dashboard. The foundation the current perps trading engine grew from.
The self-hosted backbone behind everything here: a cloud VPS running a containerized stack managed with Portainer, with every service published only through Cloudflare Zero Trust tunnels — no inbound ports exposed.
A DevOps and homelab content brand focused on lightweight single-node Kubernetes, including a commissioned platform-walkthrough video produced for Portainer.
Battle-tested technologies I ship production code with.
Open to IT leadership roles, consulting, and contract work — and always up for an interesting infrastructure challenge.