Read This First
I am not applying for a job. I'm offering an engagement.
For 30 years I've done one thing at increasing altitude: capture what only an expert can do, make it reproducible, make it self-healing, and point it at a human who needs it. In 1996 that meant a boot disk that rebuilt a hundred machines. In 2026 it means an AI software factory that takes a voice command, has one AI architect the change, another write the code, and two independent AI auditors approve it before anything ships — while I sleep.
I run a portfolio of production AI systems with real users and real revenue, built end-to-end by me. I wrote a book about how. I have a USPTO provisional patent pending (No. 64/104,528, filed July 2026) on the self-improving architecture underneath it. And I engage as a consultant or fractional executive — fixed scope, measurable outcomes, no ramp-up time, because I've already built whatever you're about to describe to me.
What I've Shipped in the Last 18 Months
Solo, end-to-end, all in production — built nights and weekends alongside my BD role until May 2026, full-time since.
Consumer visual AI
- Photo-to-cinematic-video in a customer-facing app — a real-estate customer uploads listing photos and receives an AI cinematic walkthrough video, with identity-locked characters, per-client render history, and automated quality gates. Live for paying customers.
- AI avatar / character video pipelines — persistent, identity-consistent AI characters with locked voices, natural lip-sync, and multi-scene episode assembly — finished with an automated quality review that watches every video before it publishes. It runs while I sleep, and I can run it live in front of you.
- An automated entertainment channel — episodic AI character content: script generation, render, QC gate, metadata, scheduled publishing — zero human touches from topic to YouTube. Top short: 99.4% audience retention.
- ViralStrike (viralstrike.ai) — AI video-automation platform with paying customers and per-customer provisioning.
- Sage suite (meetsage.me) — three consumer wellness apps, Google Play closed testing, live subscriptions, strict per-session privacy isolation.
The economics problem every visual-AI company has — solved in production
- Model churn: my systems have swapped generation and reasoning models repeatedly — provider deprecations included — with zero downtime, because no system of mine depends on any single vendor surviving the month.
- Cost control: hard daily spend caps enforced in code, per-call cost ledgers — and where the math favored it, migration of entire render pipelines to local GPU at $0 marginal cost.
- Quality at consumer expectations: every output passes an automated AI review gate before it ships. The gate has caught defects human review missed — including a 29-millisecond audio clip.
Enterprise AI & agents
- MADISON — multi-agent enterprise AI platform: voice executive assistant across phone, app, and web; intake classification; lead qualification; proposal generation; calendar and email tools; an autonomous marketing loop with hard spend caps and a nightly self-review. She answers a real phone number today: 605-608-9131.
- AI employees working the front line of real businesses — Ruby, Liberty, and Madison answer real phone lines for real companies today: first ring, any hour, holding conversations callers treat like a person — because that's how they're built to treat the caller. They book real appointments against real calendars, take payments, and hand off to their humans gracefully. Not chatbots with a phone number — named colleagues customers ask for.
- An evidence-based assessment engine — it measures what's actually true about a business's digital operation and publishes client-ready audit reports where every claim traces to a verified finding. "We publish nothing we could not verify" — read its real output at veloxe.ai/reviews today.
- An AI software factory — I describe a change out loud; a built, independently audited, deployed application comes back. Every change is reviewed and approved by multiple independent AI reviewers before it ships. 100+ production builds to date. This is the subject of the patent filing (mechanism under NDA).
- Self-healing pipelines as a default pattern — content filters, render failures, API outages: detect, reroute, retry with limits, alert loudly, never fail silent. My 2026 systems fail the same way my 2003 Active Directory designs did: safely, and on purpose.
A real business running on all of it — the versatility proof
- Valley View Retreat (valleyviewsd.com) — an off-grid Black Hills campground my family owns and runs, where I built everything a guest touches: the website with real-time online booking and card deposits, the AI employee who answers the phone line around the clock, and a hardened self-service kiosk — a physical device I built and deployed on-site for 24/7 guest check-in. The entire operation runs on an off-grid solar system I designed and installed myself: a 4,000-watt array and a 6 kW battery bank. The software, the hardware, and the electricity they run on — one builder. That's what hands-on means.
The platform underneath all of it
A self-hosted production estate running unattended — my own GPU compute (which is why my render costs round to zero), containerized services, full CI/CD, and web + Android + iOS delivery. Self-monitoring with automated escalation: it texts me when something needs a human, and it rarely does.
The 30 Years Before That
- Becton Dickinson (2018–2026) — Senior Software Engineer; architect and sole builder of the enterprise OSD platform — rebuilt enterprise Windows deployment for hospital-critical medical devices nationwide: a 60-page manual process and 4.5 hours of downtime per device became a fully automated 15-minute, zero-touch pipeline. Embedded automated log-intelligence into the platform — production AI in regulated healthcare infrastructure before it was fashionable. Estimated millions saved in support costs. The BD chapter closed in May 2026; the AI portfolio above — already live — became the full-time business.
- Golden West Technologies (2016–2018), Solution Architect, Rapid City SD — pre-sales through delivery: Statements of Work, storage/virtualization/HCI design, cloud migrations, training lab, junior-engineer mentoring. What I was told when they hired me: "We would never be able to afford this type of talent or get this type of talent here in Rapid City."
- Microsoft Consulting Services (2012–2016), Senior/Principal Infrastructure Consultant — Fortune 100 deployment and infrastructure design coast to coast: Veterans Administration (400,000-desktop program), Bayer (18,000 global), John Deere (18,000), Mead Johnson (8,000, led a 26-person global team), Manheim/AutoTrader (16,000), Willis (17,000). Three top-score awards for customer service and leadership.
- Before that: consolidated 40 NT4 networks into one global Active Directory forest for AOL–Time Warner; built AD + clustered messaging for the Salt Lake City Winter Olympics 5,000-athlete Cyber Village; enterprise PKI for Payless; ground-up virtualization conversions; CTO and business-owner roles. Hands-on in nearly every U.S. state.
- Certifications: MCSE (Design) · MCSA · MCITP: Virtualization · MCTS · MCP · CCNA · Foundry FCNE L2-7 · Server 2016 Compute & Storage · O365.
Recognition
"Jerry truly is a rock star. With his solid technical skills, hard work ethic, and team-player approach, Jerry helped Mead Johnson in their Win7 initiative and did a fantastic job."
— Microsoft client engagement, Mead Johnson
"I just got off the phone with a customer and he was extremely happy with Jerry. Called him a 'workhorse' and let me know that we should send him to other customer sites."
— Lisa, Dell PM, Infrastructure Services
"We could not do what we do today without Jerry's involvement, his revolutionary approach, and his commitment to making the system better every day."
— David Suarez, former Becton Dickinson program lead
The Book
I wrote it all down. The Command Layer: How One Builder Spent Thirty Years Learning to Ghost the Future (Jerry Odom, 2026 — on Amazon) is the 30-year arc of one idea — capture expertise, make it reproducible, make it self-heal — from a Ghost boot disk at Universal Studios to an AI that runs businesses.
The book is the long-form proof — the failures included, because the scars are where the lessons live. This page is the executive summary of a career spent shipping.
The Patent
USPTO Provisional Patent No. 64/104,528 (pending, filed July 2026) — covering the self-improving AI system architecture that runs my production estate: autonomous software factory patterns and dual-brain agent design. (Details available under NDA where an engagement warrants it; the running systems themselves are the public proof.)
How I Engage
Fractional CTO / Principal Architect — consulting engagements, not W-2.
- Fixed-scope builds (a platform slice, a pipeline, an AI integration) with defined deliverables.
- Fractional executive leadership: architecture ownership, model and vendor strategy, cost governance, small-team leadership — days per week, not butts-in-seats.
- Assessment engagements: evidence-backed architecture and AI-readiness audits with executive-ready reports (I run this exact engine on my own clients today).
Why consulting: you get 30 years and a working AI estate on day one, with no ramp, no relocation, no golden handcuffs. Every system named above is live: ask and I'll run any of them in front of you, in real time, on a call. I do this work every day — watching it happen is part of the pitch.
The short version: I don't talk about what AI could do for your product. I'll show you mine doing it, and then I'll build yours.
Live Links — Proof Over Promise
- viralstrike.ai — AI video platform (paying customers)
- meetsage.me — consumer wellness suite (live subscriptions)
- realtyhelm.com — AI staff for real-estate agents; call 605-608-9080 and Ruby answers. She is not a demo — she is on the job, right now, for real agents.
- alexosd.com — flagship real-estate client: a working agent whose AI employee handles her inbound today
- valleyviewsd.com — the family campground: book a site online right now, or call and the AI answers
- lakesidefamilydental.net — a complete AI front-desk build for dental practices; tap the orb and Maya picks up
- arena.veloxe.ai — three frontier models debating over live data, open-sourced (MIT)
- The Command Layer — the book