New · 2026 pricing What custom AI actually costs — three bands, read from six published guides. See the cost bands

Designed, built, and shipped by one engineer

Custom AI systems for companies that need more than ChatGPT.

AI copilots, intelligent automation, forecasting, and decision-support systems built around your data, workflows, and business goals. From strategy to deployment, I build production AI that teams actually use.

  • 8.5 years shipping production ML and AI, end to end
  • Anomaly detection across 500+ locations · 100+ live signal features
  • 3-mode AI assistant in daily use for dealer training

Trusted with production systems

M&A Supply

HVAC distribution · ~50 dealer network

Multi-dealer quoting platform replacing a third-party tool across ~50 dealers — catalog, pricing engine and proposals they own.

Dealer Training Hub

Dealer enablement · national network

3-mode AI assistant in daily use inside the training workflow. Replaced spreadsheets; leadership sees where dealers get stuck.

A live hospitality venue

Real-time audio intelligence

100+ signal features from raw room audio, running continuously and tied to operational decisions.

An investment desk

Decision support · under NDA

Instant answers over internal research and past decisions, with an auditable trail behind every recommendation.

8.5 yrsShipping production ML and AI
500+Locations under anomaly detection
~50Dealers on one quoting platform
100+Live signal features in production

Some clients are named with permission; others are described because the work sits under NDA. Every figure above is from a system running today.

Recent systems in production

Not prototypes. Systems running today, with real users and real stakes.

Financial Decision Copilot

  • Supports repeated, high-stakes investment decisions
  • Instant answers over internal research, memos, and past decisions
  • Every recommendation traceable and measured

Decision cycles got faster, with an auditable trail behind every call.

AI Dealer Training Platform

  • 3-mode AI assistant embedded into the training workflow
  • Replaced spreadsheets and manual processes
  • Real-time insight into training performance

Leadership sees where dealers get stuck and adapts material in near real-time.

Real-Time Audio Intelligence

  • Live signal processing — 100+ features from raw room audio
  • Pattern recognition with an adaptive decision engine
  • Running continuously in a live hospitality venue

A live "situational awareness" signal tied directly to operational decisions.

What problems I solve

If your team is buried in repetitive questions, manual processes, or decisions made on stale information — that's the work.

Internal knowledge assistants

Your employees find answers instantly instead of pinging the one person who knows.

AI copilots for employees

Assistants embedded in real workflows — drafting, triaging, and preparing work for human review.

Forecasting & decision support

Know what's coming with clear uncertainty bands, and decide with the numbers in front of you.

Operational automation

The "we copy-paste this 200 times a week" problems, handled end to end.

Real-time monitoring systems

Anomalies, fraud, and pattern changes caught as they happen — not in next month's report.

Custom AI products

AI as a first-class part of your product, from backend to user experience.

Why clients hire me

Most AI projects fail because they stop at prototypes. I focus on production systems:

Architecture before prompting

Measurement before hype

Evaluation built in from day one

You talk directly to the engineer building it

End-to-end ownership, data to deployment

Software that survives outside the demo

How working together works

A simple, transparent engagement model that protects both sides and keeps the project moving.

  1. 1. Scope call

    A focused call on where you are today: data, systems, constraints, and the outcome you want. Concrete terms — metrics, timelines, who owns what.

  2. 2. Fixed-phase plan

    A written plan broken into phases with explicit deliverables, success criteria, and a fixed quote.

  3. 3. Deposit & kick-off

    Work begins with a 50% deposit via PayPal business invoice to customaimodels@gmail.com. Scope locks, execution starts.

  4. 4. Weekly cadence

    We meet weekly to review progress and make decisions. You always know what's done, in progress, and next.

  5. 5. Milestone delivery & balance

    Each phase ends with the agreed deliverables: code, models, documentation, deployment. Balance due on delivery.

  6. 6. Iterate or hand off

    We move to the next phase, or I hand off clean artifacts to your team with a clear transition plan.

Ready to talk about a real project?

If you have a concrete use case and want to know what it takes to ship it, the next step is a scope call. Every quote is tied to a specific problem, phase plan, and set of deliverables.

Book a scope call Email Yeriko

Not ready for a call? Describe your use case to the assistant and get an AI-generated project assessment in under 5 minutes.