I come into a business, map how money and operations actually flow, and rebuild the parts that leak time or margin — then I ship the system myself.
Twenty years as a founder and operator: a company built and sold, a ~100-person organization led, real estate development, and a portfolio of ventures. Today I design, code, and self-host AI systems end to end — and direct a small dev team where scale needs it.
Most developers need a detailed spec to move. I don't. Give me the goal and I take it from there — prototype, design, build, deploy, measure, analyze, improve.
Tell me “homeowners should have a profitability calculator” — I don't need a 20-page requirements doc. I already understand the customer, the economics, and what “done” looks like, so I design the logic, build it, and put it in production. When the person who models the economics is the same person who ships the code, nothing gets lost between them.
The gap between an idea and a working system is usually a person who understands both sides. That's the job I do.
And when scope needs a team, I bring one: two AI-capable developers I've worked with and trust, running under my direction.
Everything below runs in production. Most of it I architected and shipped myself; two products are built by developers under my direction.
Once AI was in daily use, we realized the useful unit to pass between people was no longer a task in a task manager — it was the hand-off itself, written as a Markdown file. A product owner runs an AI working session and drops a Markdown hand-off; a designer picks it up, designs in code, and hands off HTML plus Markdown; a developer builds on it; an SEO specialist reviews and hands the developer the next requirement — again in Markdown. Every role passes structured Markdown (and the code with it) down the line, and back up it. So I rebuilt our project management around that hand-off: a self-hosted, Markdown-native workspace with Git as the source of truth, an in-browser editor with an AI coding agent pre-installed for every teammate, and a live operations dashboard — all behind a single MCP gateway (one Cloudflare Worker for sign-in and SSO, routing internal AI tools like memory, SEO, PPC, and analytics, and audit-logging every tool call for cost attribution). We moved off Basecamp onto it; the team works inside it daily.
I redesigned project management itself around AI hand-offs — and replaced our SaaS PM tool with what I built.
An always-on agent that operationalizes paid-search strategy: builds campaigns, monitors performance, defends against the platforms' auto-applied changes, and proposes fixes with one-tap approval over Telegram. CRM-grounded, governed by explicit written rules, with every action logged.
Replaces the day-to-day work of a PPC manager — with a human approval gate on every change.
kpicreatives.com/hunterA self-hosted agent exposing dozens of tools across analytics, discovery/SEO, channel management, captions, and bulk reporting — multi-channel, queryable in plain language through an AI client.
Replaces a stack of dashboards and manual reporting with one agent you can just ask.
kpicreatives.com/oscarI bought a server, set up Linux, and installed Coolify — then moved recurring vendor tools onto self-hosted open-source I selected, deployed, and integrated: Frame.io → FreeFrame (media review), DocuSign → Documenso (e-signatures), Attio / GoHighLevel → a self-hosted Twenty CRM, Calendly → Cal.com (scheduling). I run the whole layer — Docker, reverse proxy, TLS, backups — plus invoicing automation wired into Mercury.
Cuts recurring subscription spend and keeps the data in-house.
A real-estate decision platform I built alone: the site on a ~1,400-line information architecture I authored (category design, programmatic pages, structured data, AI-search readiness) — plus the full acquisition machine behind it: funnel, voice AI agent answering inbound calls, CRM integration, and analytics. Business thesis to production, no handoffs.
Positioning, funnel, voice agent, CRM, code, deploy — one person, working in production.
aduscale.comAn AI that actually watches and listens to a video — reading the footage, audio, pacing, and hooks against patterns of what performs — and tells you, before you publish, whether it will land with your target audience and travel on social. Under the hood it grades the edit against two rulebooks at once: the brand's (brand book, tone of voice, brief) and the mechanics of viral performance. You get a clear score, a director's note on what's weak, and a timecoded list of exactly what to fix — plus a quality seal to hand the client once it passes. I own the product and architecture; a developer builds under my lead.
Know whether a video will land with its audience — before you publish — and exactly what to fix.
directormode.aiA concierge agent for general contractors that files and schedules building inspections with the LA Department of Building and Safety (LADBS) — and, underneath, a compliance engine that scores every contractor on how they actually perform: who passes inspection on the first try, who needs a second or third, and with which violations. It runs on ~11 million data points covering every LA inspection over the past 13 years, so it separates strong contractors from weak ones from the record, not the reputation.
11M public inspection records turned into a contractor-quality signal — who to trust with the work, backed by data.
inspectpilot.aiI'm my own DevOps — from DNS and TLS to CI/CD, containers, databases, and observability.
Real estate is my second domain: VP of Real Estate Development, then founder of a real-estate decision platform. Here is how the systems above map onto the cost structure of a residential operator — each line is a pattern already running in production, not a proposal.
In residential rentals, marketing, broker fees, and sales commissions can eat half the cost base. That is the line Hunter was built to attack: autonomous paid-search with a human approval gate, plus SEO/AEO and full-funnel analytics — every acquisition dollar visible and defended.
When landlords, partners, and banks make refinancing decisions off your reports, the speed and reliability of the financial block is the product. I build real-time, audit-ready reporting pipelines instead of month-end scrambles.
Every acquisition leaves behind systems that don't talk to each other. I've done the unglamorous work of merging tools into one source of truth: self-hosted CRM, a unified data layer, a single authenticated gateway in front of it all.
Thousands of support tickets a month follow a finite set of scenarios — the sweet spot for AI ticketing. I run the pieces this needs in production: a voice AI agent (Vapi) that works both directions — answering residents and calling vendors to schedule repairs — wired into the CRM so every ticket, contact, and action lives in one source of truth. AI triages and closes the routine, the voice agent makes the calls, and humans approve what matters.
AI made forging an application package cheap; the counter is structured data, pattern detection over history, and a human review gate. That's the exact shape of InspectPilot — a compliance engine that scores actors from millions of historical inspection records — and the same approach transfers to screening applicants and vetting the maintenance contractors you hire.
Led a ~100-person company post-acquisition with full P&L responsibility; built and sold a business of my own. Fluent in margins, cost structure, and the numbers behind decisions.
VP of Real Estate Development at a construction holding, and co-founder & CEO of a platform that models build-or-don't ROI before a dollar is spent. Comfortable with capex, project budgets, and multi-unit economics.
Two decades turning chaotic operations into documented, repeatable systems that run without me — and now automating them with AI so lean teams outperform large headcount.
I attack recurring spend and manual work directly: self-hosting over per-seat SaaS, agents over headcount, real-time reporting over month-end scrambles.
I turn procurement into a data step, not a guessing game: send the same RFQ to a vetted set of contractors, normalize their bids to compare like-for-like (bid leveling), and score each vendor on compliance and track record — the same engine idea behind InspectPilot, which ranks contractors on ~11M inspection records. Fewer bad hires, sharper prices, and a defensible paper trail for owners and lenders.
Co-founded and led a real-estate decision platform; built the product solo — site, information architecture, and the full acquisition machine (funnel, voice AI agent on inbound calls, CRM, analytics). Paused Jul 2026.
Build and run lean, systemized operations across several ventures. Design the financial and operational infrastructure, automate the repetitive work with AI, and keep headcount low by design — while directing a small dev team.
Held a C-level role at an IT holding and served as VP of Real Estate Development at a construction holding. Owned budgets and capex decisions, and founded a real-estate financial-modeling platform.
Startups, events, and consulting across industries — a range of ventures between the agency exit and my next operating roles.
Built a digital agency from zero and sold it to a major advertising holding (first exit). Led the merged ~100-person company — global brand clients — with full P&L ownership. Won an industry effectiveness award.
Built a press office from scratch and ran editorial teams. Learned to stand up an operation with no template and make it work under pressure.
I come in, map the financial and operational reality, find the leaks, install the system, and hand it back running.