How do you price work?

I map the workflow and agree on the scope first. You get the plan, delivery points, and quote before I write code.

You own the code, accounts, and infrastructure. Handover is part of the job.

What if I do not know whether I need software, automation, or AI?

That is fine. I start with the people, steps, and result involved. I choose the technology afterward. Sometimes the right answer requires no custom code.

Can you work with my existing team?

Yes. I can work alongside your team or build the system and hand it over.

What happens after launch?

I can continue to run and improve it, or hand it over with documentation.

When does custom software make sense?

Build it when existing tools force workarounds or duplicate entry. It also fits when off-the-shelf software cannot follow a core process. If an existing tool fits, use it.

Taking on new projects

Fix what keeps slowing your business down.

I map the workflow with you before choosing the answer. We may keep your tools, connect them, build custom software, or apply AI. You get a plan and quote before I write code.

Tell me what's stuck Chat on

Based in: Vietnam · GMT+7
Works in: English · Tiếng Việt
Replies within: One working day
Now building: Reupmatic · PixelBid

Products

What I have built

4 projects, each with a public link you can open.

Reupmatic

A desktop app that processes Douyin videos in batches on your machine.

Desktop app2026Building

Visit site

The problem: Each video repeats the same download, edit, subtitle, and voice steps.
What I built: A local queue that runs those steps across a full batch.
Where it stands: Alpha builds run on macOS and Windows. I am adding direct publishing next.

AI Engineering Atlas

An AI engineering roadmap that starts with what the learner is missing.

Web product2026Live

The problem: Working developers know AI unevenly, while most roadmaps restart from zero.
What I built: A diagnostic that finds the gaps and creates the shortest route to a practical project.
Where it stands: The source code is public, and the website is live. I continue to expand the material.

copycat-skills

Check the market before deciding whether to build the app.

Agent skill2026Shipped

The problem: One product link is not enough evidence for a market decision.
What I built: An agent skill that finds competitors, demand signals, and reasons to build, stop, or change direction.
Where it stands: You can install it from skills.sh and inspect the source on GitHub.

PixelBid

A working rebuild of a bidding game that gained traction on X.

Web product2026Building

Visit site

The problem: I wanted to understand the auction mechanic instead of guessing from the interface.
What I built: A credit-backed attention market: a qualified open burns one credit.
Where it stands: I am building it as a side project. The original idea came from outbid.lol.

Digital operations

When custom software is worth it

AI for repetitive desk work
The signal
Your team reads, sorts, summarizes, or answers the same type of information each day.
What to change
AI handles repeated steps inside your current systems. A person approves high-stakes decisions.
Move field work into one system
The signal
Staff receive jobs, record work, or report progress through paper, chat, and spreadsheets.
What to change
A mobile app built around the workflow that still works on a weak connection.
Connect a fragmented workflow
The signal
Data gets copied between files, chat threads, or systems that do not share it.
What to change
Internal software that connects the steps, cuts re-entry, and shows where each job stands.
Test a new digital product
The signal
You have a new service or revenue idea but do not yet know whether customers will use it.
What to change
A small SaaS product that customers can use and pay for before you expand it.

From bottleneck to operation

A plan only matters once your team can use it.

Choose a launch point and see when the work starts creating value.

1Locate the bottleneck
Who does the work, where it stalls, what it costs.
Not in use yet
2Choose the smallest useful fix
Change the process, connect tools, or write code.
Not in use yet
3Run an early version
Your team tests it with real data.
Not in use yet
4Put it into operation
Measure, fix, and hand over.
In use

Put it into use

A small first releaseDuring the buildAfter acceptanceNever put it to work

I finish the agreed workflow before putting it into operation. Then measure, adjust, and improve it.

Ways to work together

Start small. Put it to work.

Depending on the problem, I can stop at a clear plan, test one workflow, or take the full system into stable operation.

01
Find the bottleneck
The outcome
The current workflow, solution options, the smallest useful scope, a plan, and a quote.
Use this when
Work is slow, but you do not yet know whether the process or the software is at fault.
02
Put one workflow into use
The outcome
One important workflow running with your team and real data.
Use this when
You want evidence from daily use before expanding the system.
03
Build the full system
The outcome
Software in operation, with documentation, handover, and a maintenance plan.
Use this when
You want one person responsible from workflow mapping through stable operation.

Before we start

What to settle before the work starts

Pricing, ownership, and choosing what to change.

How do you price work?: I map the workflow and agree on the scope first. You get the plan, delivery points, and quote before I write code.
Who owns the code?: You own the code, accounts, and infrastructure. Handover is part of the job.
What if I do not know whether I need software, automation, or AI?: That is fine. I start with the people, steps, and result involved. I choose the technology afterward. Sometimes the right answer requires no custom code.
Can you work with my existing team?: Yes. I can work alongside your team or build the system and hand it over.
What happens after launch?: I can continue to run and improve it, or hand it over with documentation.
When does custom software make sense?: Build it when existing tools force workarounds or duplicate entry. It also fits when off-the-shelf software cannot follow a core process. If an existing tool fits, use it.

Start with the work that is stuck, not an app idea.

Tell me which step is slow, error-prone, or consuming too much staff time. I will map it with you, then send a plan and quote. If you do not need custom code, I will say so.

Tell me what's stuck Chat on

Fix what keeps slowing your business down.

Reupmatic

AI Engineering Atlas

copycat-skills

PixelBid

AI for repetitive desk work

Move field work into one system

Connect a fragmented workflow

Test a new digital product

Find the bottleneck

Put one workflow into use

Build the full system

Start with the work that is stuck, not an app idea.