Portfolio
AI & Automation
I build systems that replace manual workflows with AI-powered automation. From no-code platforms handling thousands of transactions to autonomous AI agents that research, write, and execute with minimal human oversight.
Washington State / 14 years, U.S. Air Force / MBA, University of Washington
Selected Projects
n8n pipeline that does what Clay does, built on client infrastructure for 12 B2B engagements
Built and operated AI-powered outbound infrastructure for 12 B2B clients across SaaS, agencies, events, and professional services. The core was an n8n pipeline that did what Clay does, deep-enriching every prospect, but built entirely on the client's own stack. A deep research agent node pulls LinkedIn data, company news, and website content to assemble a context packet for each lead. An AI writing node turns that packet into a one-line icebreaker, a tailored email body, and a subject line. Replaced 15 minutes of manual prospect research per lead with a 30-second agent run. Integrated client systems end to end via REST APIs and webhooks (Salesforce, HubSpot, Airtable, Apollo, LinkedIn Sales Navigator) and owned the full deliverability stack (SPF/DKIM/DMARC, warmup, bounce tracking) through Instantly.
Multi-agent Claude Code pipeline running 24/7, with human approval in Slack
Built for a YouTube client averaging 500–1,000 views per video. A persistent Claude Code agent on a 24/7 VPS runs the entire script pipeline, orchestrating specialized subagents with structured logging, error handling, and monitoring throughout. Each week it pulls the top 5–10% of performing videos from the client's channel and competitors, extracts the patterns, scrapes comment sentiment for what's resonating or missing, researches the specific topic, and layers the channel's own worldview on top. It generates an outline and pauses, pushing it to Slack for a human reviewer who can approve or redirect in plain English. Because it's Claude Code and not a rigid workflow tool, the reviewer can say "make this less hype-y, lean into the tax angle" and the agent rewrites accordingly. Once approved, it drafts the full script using the client's tone, length targets, humor, and story patterns from a detailed CLAUDE.md. A hook agent generates 100 variations, scores each against a rubric, then applies a Karpathy-style ratchet (single-word mutations) to push the top hooks higher. Subscribe and like CTAs run through the same treatment. The finished script is reviewed by two separate agents, one for guideline compliance and one for factual accuracy plus math verification, before a human does the final read. Weekly, the system pulls transcripts and performance data, identifies what's working, and tunes the rubrics and instructions.
Full-stack no-code platform built in Bubble.io for a flat-fee brokerage
A flat-fee real estate brokerage was running hundreds of simultaneous transactions on spreadsheets, email threads, and manual status tracking. Deals got lost, documents scattered, and there was no single view of operations. I designed and built the full platform on Bubble.io: database architecture, role-based auth, frontend, and UX. That replaced their entire manual workflow. It handles listing intake, state-specific compliance tracking, task management, agent assignment, escrow coordination, payment processing, and revenue reporting. Integrated Stripe for payments, DocuSign for contract e-signatures, Zillow for listing data, and OpenAI for automated listing-description generation. In the first 18 months, 1,900+ homeowners listed properties through it across four states.
Background
14 years as an officer taught me how to break complex operations into repeatable processes. That mindset translates directly into automation: identify the workflow, find the failure points, build the system that eliminates both.
University of Washington MBA. I speak both business and technical fluently. I can sit with a VP of Sales and understand their workflow, then go build the automation that fixes it. That bridge is the hard part, not the tooling.
I started my career as a medical laboratory officer running statistical QC programs: regression analysis, control charts, and HL7 data pipelines routing 742K+ test results into Cerner every year. That's where I learned to build systems that monitor their own accuracy. It's why every AI workflow I ship has structured logging, fact-check agents, and feedback loops instead of hope.
As a GTM Engineer I built outbound infrastructure across 12 B2B client engagements: 576K+ emails sent, 4,365 positive replies, and 1,784 qualified meetings booked with decision-makers at Google, Uber, Vimeo, and Stripe. Nothing in this portfolio is theoretical. It's all in production, producing revenue for real companies.
I'm looking for a remote role where I can build AI-powered automation systems that make teams faster and operations smarter.