The Launch Review

You built an app that works.
Now people are going to use it.

The engineering read you’d get from a technical co-founder — for founders who don’t have one.

Book the review Read the checklist first

$1,500
Fixed price: One week
Start to finish: 2.25 hrs
Of your time

Sound familiar?

You’ve been asking these. Nobody has answered them.

Why is my app slow?

Why is this deploy failing?

Which cloud should I use?

What is this actually going to cost me?

Why are there random bugs I can’t reproduce?

What do I do before real users show up?

You get all six answered — about your app, with your numbers.

What you get

Four things. The first one is a picture of your own system.

Start here

Your system, drawn as one diagram

What that’s worth

Every piece of what you’ve built and how the pieces talk to each other, on a single page, labelled in English. Most founders in this position have never seen one. It’s also the first thing you’ll send the first engineer you hire.

Illustrative — yours has your actual pieces on it

Straight answers to the questions you arrived with

The ones you wrote down before we spoke — answered with actual numbers, not “it depends”. What it costs today, what it costs at ten times the traffic, and where it stops working.

Two ranked lists: fix now, and don’t build yet

In order, not as a pile. Fix now is what stands between you and real users. Don’t build yet is the expensive work you’ve been told you need and don’t — with the condition that would change my mind.

The readout call, recorded and yours to keep

Ninety minutes walking through all of it, recorded. Watch it again at your own pace, or hand it to whoever joins next instead of explaining it yourself.

A week of async questions afterwards, included. The questions worth asking are the ones you think of two days later.

What I’ll tell you not to build

Half of this is talking you out of things.

Kubernetes

You probably don’t need Kubernetes.

Why not

It's the default answer from people who run it for a living. For one app and one database it's weeks of work to end up somewhere slower to change than where you started.

When that changes

When you're running enough separate services that starting and stopping them by hand is somebody's actual job — or when a customer's contract requires it.

Microservices

You probably don’t need microservices.

Why not

Splitting one working app into six means six things to deploy, six things to watch, and network calls where you used to have function calls. The complexity lands on day one; the benefit arrives much later, if ever.

When that changes

When separate teams are blocking each other in one codebase, or one part of the app genuinely needs to scale independently of the rest.

Leaving Vercel

You probably don’t need to leave Vercel.

Why not

Moving to raw cloud infrastructure to save money usually trades a bill you understand for a bill you don't, plus the ongoing job of running it. At your traffic, the platform is cheaper than the hours.

When that changes

When the platform bill is genuinely material against your runway, or you hit a hard limit — long-running jobs, a region you need, a compliance requirement.

Nobody else is paid to tell you this. That is most of what you’re buying.

How the week goes

Two calls from you. The rest is mine.

45 min

We talk

You tell me what you built and what worries you.

Bring the questions you've been carrying around. No preparation, no deck, nothing you need to understand first — if you could explain it already you wouldn't need the call.

A few days

I read

Your code, hosting, database, deploys and bills.

Nothing needed from you while it happens. I don't change anything and I don't need you on hand — if something's unclear I'll ask in a message.

90 min

We go through it

The diagram, your answers, and the two lists.

Recorded, so you can watch it again instead of taking notes. A week of async questions afterwards, included.

One week from the first call to the recorded readout. No production passwords, and I don’t change anything.

Fit

Who this is for, and who it isn’t.

Who it’s for

You built it with AI tools.Claude Code, Cursor, Lovable, Replit — the product exists because you didn’t wait for an engineer.
People are using it, or about to.Real users, real money moving, or a launch date already on the calendar.
There’s no engineer to ask.No CTO, no technical co-founder, and nobody whose job it is to tell you what’s true.

Who it’s not for

You have an engineering team and a real platform.You don’t need this translated. You need the infrastructure teardown.
You want me to build it for you.This is a read and a plan. Doing the work is separate, and quoted after.
Nothing is deployed yet.There’s no system to look at until it’s running somewhere and someone is using it.

You have engineers?Already have an engineering team and a real platform?Then you don’t need the translation layer. The infrastructure teardown is the same work, written for people who already speak the vocabulary.See the infrastructure teardown

Cost

One number now, and the next one before you ask.

The review$1,500

Fixed. Agreed before we start.

Not an estimate and not an hourly rate that grows once I've seen the codebase. One week from the first call to the recorded readout. If I read your setup and there's genuinely nothing here worth your money, I'll tell you on the first call, before you've paid anything.

If you want it done too$3,500–$6,000

The $1,500 is credited in full against it.

If you want the fix-now list done rather than done by you, I quote it after the review — from a list you’ve already read, so the number doesn’t move afterwards.

$1,500/monthOptional afterwards: a few hours a month to be the engineer you ask before you decide. Cancel any time.

Who I am

Who’s actually reading your app.

Talha Imtiaz

Software engineer

I write code first. That’s why I can read what you built, not just judge where it’s hosted.

The longer version

I spent two years writing application code before I spent the last year keeping other people’s systems running. Most of my work is the same job this page describes: someone shipped something that works, more people arrived than it was built for, and it needed to hold. You work with me directly, start to finish — there is no team behind this and nobody else will be on your calls.

~80 → 400+concurrent users on gkeRead how

How that was done

KEDA scaling on a waiting-room player threshold rather than CPU, with API and Hocuspocus split into separate pods.

I write these the same way I write them for engineering teams. I just don’t assume you speak the vocabulary. The findings are the same findings.

Before booking

What people ask before they book.

01Do you need access to my code?

Yes — read access, and nothing that lets me change anything.

Your repository, and read access to wherever the app runs and stores data. I don’t need admin, I don’t need to be able to deploy, and I don’t touch anything — I’m reading, not editing. If you’d rather not grant access at all, we can do it as a screen share instead; it takes longer and you have to be there for it, so it’s the fallback rather than the default.

02What if you don’t find anything serious?

Then that’s the answer, and you get it in writing with what I checked.

It’s a genuinely useful thing to know, and it’s the thing you’d otherwise pay to find out the hard way. I won’t invent problems to justify the invoice — if the honest verdict is “this holds, here are the three things to watch”, that’s the readout, and you still get the diagram and the answers.

03What if I don’t understand the recommendations?

Then I wrote them wrong, and I’ll rewrite them.

That’s the failure this whole review is designed against — a document written for engineers you don’t have is worth nothing to you. The readout is 90 minutes and recorded so you can watch it again at your own pace, and a week of async questions afterwards, included. Ask the same question three times if you need to; that’s what the week is for.

04Do I have to hire you afterwards?

No, and it’s never a condition.

Most of what comes out of this is work you or a contractor can do from the list. If you’d rather I did it, I quote that separately after you’ve read the findings — $3,500–$6,000, with the review fee credited in full. Nothing about the review changes based on which you pick.

05Will you sign an NDA?

Yes — send yours and I’ll sign it as-is.

Before I get access, ideally. If a clause is genuinely unusual I’ll tell you which one and why rather than quietly not signing. If you don’t have one, a standard mutual NDA is fine.

06How do I pay?

50% at booking, 50% on delivery.

You'll get an invoice with a card or US bank-transfer link — the same way you'd pay any US vendor. Prefer Payoneer, Wise, or direct wire? Those all work too; just say so and I'll send details. You’ll get a proper invoice either way — this isn’t a transfer to a stranger, and it’s the kind of receipt your accountant expects.

Next step

Book the review.

The first 45 minutes cost nothing. If this isn’t worth your money yet, I’ll say so on that call.

Your system, drawn as one diagram
Answers to your questions, with numbers
Fix now and don’t build yet, both ranked
90-minute recorded readout, yours to keep

Price$1,500Fixed, agreed before we start.

TimeOne week

Book the review Read the checklist first

Or email hello@talhaimtiaz.me