Talk to me

VAI Soft Lab · Agentic AI architect · Sydney

I design and build agentic systems that deliver enterprise software.

A one-person studio that uses AI fully and openly: agent systems do the heavy work, I design, lead and review. From an idea to a full app, a build, fix or finish of your app, agents and automation inside your team, or a custom AI app for your daily work. I lead every project myself, from start to finish.

Have an app idea, a stuck app, or a daily job you want an AI agent to do? Write three lines. I reply within one working day with a plain answer.

Sydney, Australia. Clients in Australia and overseas.

Vahid Alamipour, founder of VAI Soft Lab
Vahid AlamipourAgentic AI architect, founder

How I can help

Three ways to work with me.

I join your team.

As your lead engineer, paid by the hour. I own the architecture, bring agents and automation into how your team builds, ship features and raise the quality of the whole app. When it runs on its own, I leave.

Paid by the hour

I build, fix or finish your app.

From an idea to a full app, an AI feature inside your product, or a half-built app that needs finishing (including apps built with AI tools). First we define the scope together: the job, how I help and how long it takes. Then you get the time and the price. Any technology you want.

Scope first, then time and price

Custom AI apps for your daily work.

Your team does the same computer work every day: quotes, orders, invoices, reports, copying the same numbers from one program to another. I sit with the people who do the work and learn how it is really done. Then I build a small AI app that does the routine part, while your staff approve the important steps. You know your business. I know software and AI.

Study first, then a fixed price

Prices in Australian dollars (A$).

Questions

Questions about the AI apps.

How does it start?

With a short Study at a fixed price. It takes one to two weeks. I come to your business, or join online, and sit with the people who do the work on a normal day. At the end you get four things: a one-page map of your daily work, the one job we should do first, a small working first version of the app, and a fixed price for the first app. Then you decide.

What kind of work can the app do?

Routine computer work that follows clear rules and happens every day. Some examples: turning a customer's email or a photo of a form into a draft quote, typing an order into your system, chasing missing paperwork, writing the weekly report from numbers you already have, or moving the same data between two programs. The app never moves money, and it never makes big decisions about people on its own, for example about their health, a loan, a job or a home. If a person should decide, a person decides.

What if the AI makes a mistake?

It can happen, like with a new staff member, so I build the app the safe way. Before anything runs, we write down together what the app may do on its own, what needs a person's OK, and what it must never do. The app starts in "draft only" mode: it prepares the work, and a person checks it before it goes out. Then it runs next to your staff for a few weeks before it goes live. Every action is logged, and there is an off switch. The important steps always stay with your people.

Where does our data go?

It stays yours. The app runs in your own accounts, not mine. For personal information, such as customer names and addresses, the AI runs in a business cloud service in an Australian data centre, with the account in your name. Your data never goes into a consumer AI tool. If the app talks to your customers, it tells them it is an AI.

Do you need to know my industry?

No. You know your business. I know software and AI. Every business does its work a little differently, so I learn yours by watching the real work, not by reading about your industry. That is what the Study is for.

What happens after the app is live?

Two choices. You keep me on a monthly plan, which I call Care: I keep the app running, fix small things, and make small improvements as your work changes. Or you take the app and run it yourself. The code is yours either way, and I hand it over with the documentation.

How it works

From your notes and documents to the real app.

Two steps. VAI Cortex turns everything you give me into a blueprint you agree to. VAI Development, my agent team led by me, builds the real app from it. I lead every project myself, from start to finish.

01 · You give me
  • Meeting notes
  • Documents
  • Requirements
  • Slack messages
  • Emails
  • PDFs
  • and more…
02

VAI Cortex

Collects everything about your project. Keeps learning as we go.

03
The blueprint
  • Brief
  • Requirement pack
  • Design

Brief, requirements, design

Cortex writes it. You read it in plain language.

04

You agree

Nothing is built before this.

05
VAI Development
Vahidhuman · leads the team
PM agentruns the pipeline
Designer
Lead engineer
QA
Developer
Developer
Developer

VAI Development

My agent team builds it. I lead it.

06

The real app

  • Code quality
  • Easy to maintain
  • Accurate
  • Secure

VAI Cortex: collect and understand.

Everything about your project goes in: meeting notes, documents, requirements, Slack messages, emails, PDFs. Cortex keeps learning as the project goes on, and turns it all into a brief, a requirement pack and a design: the blueprint. You read it, you agree, and only then we build. No surprises later.

VAI Development: build the real thing.

My agent team turns the agreed blueprint into the real app. I lead it. A project manager agent runs the pipeline; designer, lead engineer and QA agents shape the work; developer agents write the code, the tests and the documentation. I design the solution, lead the project from start to finish, run the tests and try the app myself. You get lead-engineer quality, faster and for less.

  • Code qualityClean code your own team can read and extend.
  • Easy to maintainClear structure, tests and documentation, so changes stay cheap.
  • AccurateBuilt from the blueprint you agreed to, and checked against it.
  • SecureYour code is yours. Your data never goes into a consumer AI tool. Security is part of every review.

VShell

One of my tools: VShell.

VShell is a Mac app I built. You talk only to an AI project manager, and it runs a team of AI agents on your project. Claude Code, Codex and Cursor in one window, and prompts by voice.

Apple chip Mac, macOS 14 or newer. AI use not included.

VShell team page: an AI project manager and AI member cards, the usage of each account and a "Needs you" list.

Our games

And a game: Hold the Home.

Hold the Home is a free zombie defense game made by VAI Soft Lab. Your phone is the controller and the map. Your TV or computer shows the 3D battle. 1 to 4 players, no install.

In your web browser. You need a phone and a TV or computer.

Hold the Home banner: hands hold a phone with the game map, and a big TV shows the battle at the house in the forest.

Proof

Proof

0+years building software
0+developers led into AI engineering
0Anthropic certificates in building with Claude
0%of my projects led by me, start to finish
  • VAI Cortex + VAI Developmentmy own agent system, from your notes to the real app. See how it works, or ask me for a live demo.
  • Tools I buildfor example VShell, a Mac app where one AI project manager runs a team of AI agents. See it. And games, like Hold the Home.
  • Enterprise appsticketing, healthcare, financial services, property and B2B software, live on the web, iOS and Android.

About me

I'm Vahid Alamipour.

16+ years building software: from junior developer to CTO, then lead engineer roles in Malaysia and Australia. Today I am an agentic AI architect and lead AI engineer in Sydney: I have led 30+ developers into AI engineering, and I design the agent pipelines that turn requirements into working apps. Over a year of daily Claude Code work. Four Anthropic certificates in building with Claude.

VAI Soft Lab is a one-person studio. I use AI fully and I say so: agent systems do the heavy work, I design, lead and review, and I take full responsibility for what I ship. That is how you get best-practice software of the highest quality, faster and for less.

I work in any technology you want. My deepest hands-on stack is React, React Native, Next.js, TypeScript and Node, with .NET, GraphQL, PostgreSQL and AWS behind it. Domains I know well: ticketing, finance and loans, healthcare, property, construction, B2B software, user onboarding. I work in English, in the Sydney time zone.

  • Agentic AI
  • Claude Code
  • OpenAI Codex
  • TypeScript
  • React
  • React Native
  • Next.js
  • Node
  • AWS
  1. StartJunior developer
  2. ThenCTO
  3. MalaysiaLead engineer
  4. AustraliaLead AI engineer
  5. NowVAI Soft Lab, Sydney

Contact

Have an app idea, a stuck app, or a daily job you want an AI agent to do?

Write three lines. I reply within one working day with a plain answer.

  • Sydney, Australia
  • ABN 53 361 734 507