StackProofStackProof

For agencies

Anyone can build the demo. Very few can build the thing that still runs in six months.

If you sell AI automation, you already know the failure: it works in the demo, it runs for two weeks, then it stops quietly at 2am and your client is the one who notices. You eat the hours, and you lose the retainer anyway.

I am the engineer agencies bring in so that does not happen. White-label, fixed price, under your brand. My background is production AI that runs unattended at industrial sites, where nobody is watching the screen and an operator calls when it stops. That is a different craft from wiring a happy path, and it is the entire reason this page exists.

Automation rescue

$900 fixed · 5 business days

Something you shipped is fragile or already broken.

  • ·Written audit of every failure mode found, ranked by how it bites
  • ·The fixes applied: retries, backoff, idempotency, real error handling
  • ·Alerting, so you hear it from a monitor and not from your client
  • ·A one-page runbook your team can act on without me

Start here. It is a cheap way to find out if I am any good before you trust me with a build.

Production build

from $2,500 fixed · about 2 weeks

You sold it. I build it, under your brand.

  • ·The working system, scoped and quoted in writing before anything starts
  • ·Structured logging on every external call, so failures are explainable
  • ·Monitoring and alerts wired to wherever your team actually looks
  • ·Handover documentation good enough that you could replace me

Your client never has to know I exist. Code, docs and infrastructure are yours.

Retained capacity

$2,800 per month · ongoing

You need an engineer, but not a salary.

  • ·A standing engineering slot: roughly one build plus maintenance monthly
  • ·Priority queue ahead of project work, and a direct channel
  • ·I join client calls under your brand when it helps you close
  • ·Continuous care of everything already shipped

Month to month. If a quarter goes by without it earning its keep, cancel it.

What ships with every engagement

This list is the whole difference. It is unglamorous, it is what separates a system from a prototype, and it is why the second project tends to follow the first.

  1. Structured logs, with a record of every external call
  2. Retries with backoff on anything touching a network
  3. Alerting on failure, routed where your team already looks
  4. A runbook: what it does, how to restart it, what breaks first
  5. Handover docs written so your team is never trapped
  6. Fixed scope in writing, with the exclusions stated up front

Proof you can click right now

The site you are reading is one of the exhibits. It is not a brochure: it is a running system that maintains itself.

Behind that sits the work I cannot show you: multi-camera AI detection running around the clock across distributed industrial sites, with fleet operations, hardware failure diagnosis and written incident runbooks. I do not name clients or show their data, which is the same discretion you would want from anyone you let near yours.

Where I am, plainly

Vietnam, UTC+7. I lead with that rather than letting you discover it later. My mornings overlap your evenings, and handover documentation ships with everything precisely so a time zone is never the thing standing between you and your client. You are paying for the engineering, not for an office in your city.

How to start

  1. Email two lines: what the automation does, and what it does wrong.
  2. You get a fixed quote and a written scope within 24 hours. No call needed.
  3. Half up front, half on delivery. Everything produced is yours.

lienhe@mmct.com.vn

A reply of "we are covered, thanks" is genuinely useful too, and I will not chase.