How we work
How a project runs, week by week
One method, a weekly rhythm and a few promises that hold on every project. Here's what working with us looks like, from the first call to life after launch.
Habits
Three habits behind every project
-
The people you meet build it
The people you meet on the first call are the people who write your code. No hand-offs to a junior bench.
-
AI in every workflow
We use AI to speed up research, testing and review, so you get more shipped for the same budget.
-
Shipping every week
You see working software every week, not a status deck. Small releases keep risk low and feedback fast.
Process
The Alpha Method
Every project runs through the same four stages: map, prove, ship and scale. Each one ends in something you can open and check.
-
Map
We learn your goals, users and constraints, then agree on what success looks like.
- Scope document
- User flows
- Risk list
-
Prove
We design and test the riskiest part first, so big decisions rest on evidence.
- Clickable prototype
- Technical spike
- Estimate
-
Ship
We build in weekly releases with demos you can click, not slides.
- Working releases
- Test suite
- Staging environment
-
Scale
After launch we monitor, improve and grow the product with you.
- Monitoring
- Roadmap
- Support plan
Engagement
Ways to work together
Every project starts the same way. After that, you choose how to pay for the build and what happens after launch.
Discovery sprint
Fixed price. Ends with a written estimate for the build.
Launch
Your product goes live.
Cadence
A week with us
However far apart we are, every week runs the same way.
Three hours of overlap
At least three hours of our day line up with yours.
A written update every day
We post a written update daily, wherever you are.
A demo every week
Working software you can click, not a status deck.
A reply within one business day
Any message, before and during a project.
AI
Where AI helps, and where people decide
People make the decisions and own the judgement calls. AI speeds up research, testing and review.
-
Research
AI speeds up the research.
People make the decisions.
-
Testing
AI speeds up the testing.
People own the judgement calls.
-
Review
AI speeds up the review.
People make the call.
- Your data stays in accounts you control.
- Model providers' business tiers, which don't train on your data.
- Built so the model can be swapped later.
Every project
What you always get
These hold on every project, whatever its size or how we price it.
-
You own the code
Code, designs, prompts, test sets and infrastructure are yours from the first commit, in repositories and cloud accounts you control.
-
Weekly demos
You see working software every week, in a demo you can click, not a status deck.
-
Three hours of overlap with your team
However far apart we are, at least three hours of our day line up with yours, and we post written updates daily.
-
A reply within one business day
Before and during a project, any message gets an answer within one business day.
Stack
Proven tools, chosen for the job
Drawn as one request travels: from the screen in your hand, through the backend and AI, down to the cloud it runs on.
-
Web and mobile
Frontend
- React
- Next.js
- Vue
- Tailwind CSS
Mobile
- Flutter
- React Native
- Swift
- Kotlin
-
Backend
- Python
- Django
- Node.js
- PostgreSQL
-
AI
- OpenAI
- Claude
- LangChain
- pgvector
-
Cloud & DevOps
- AWS
- Docker
- Kubernetes
- Terraform
- GitHub Actions
FAQ
Questions we answer before you hire us
Straight answers to what founders and product teams ask us before they hire us.
Still deciding?
Ask the people who would build it. We reply within one business day.
Cost and timeline
Every project starts with a fixed-price discovery sprint. It ends with a written estimate for the build, priced as fixed milestones or as a monthly team, so you see the full number before committing to anything beyond the sprint.
Usually 6 to 12 weeks. The discovery sprint fixes the scope first, so the timeline is agreed before development starts, and you see working software in a demo every week after that.
They usually do, and that's fine. We estimate each change in writing before building it, so you choose whether to swap scope, add time or hold the release.
AI projects
RAG (retrieval-augmented generation) lets an AI model answer from your own documents and data, and show its sources. You need it when answers must be accurate, current and specific to your business.
We agree a test set and a pass mark with you before we build. The Prove stage measures the riskiest part against it, and if it misses, you decide whether we adjust the approach or stop. You keep everything built up to that point.
Yes. Your data stays in accounts you control, we use model providers' business tiers that don't train on your data, and only the people on your project can access it.
Whichever fits the job: Claude, OpenAI models or open-source models you host yourself. We build so the model can be swapped later without rewriting the product.
Working together
Yes. We turn business goals into a technical plan, explain trade-offs in plain language and handle the engineering end to end.
We keep at least three hours of overlap with your team, post written updates daily and demo working software every week.
Yes, including prototypes built with tools like Lovable or Replit. We start with a short audit of code quality, security and infrastructure, then propose a plan to stabilize and extend it.
Ownership and support
You do. Code, designs, prompts, test sets and infrastructure are yours from the first commit, in repositories and cloud accounts you control.
We stay on to monitor, fix and improve the product on a monthly support plan, or hand over with documentation if your own team takes it on.
Contact
Have an idea or a stuck project? Let's talk.
We reply within one business day. If we're not the right team, we'll say so on the first call.