Product engineering studio

Plan, design, build, run. One studio.

No handoffs, no account managers. One engineer owns your product from the first call through to handover. Fractional CTO when you need direction, hands-on code when you need shipping.

Web products and AI systems, for founders and established teams alike. UK-based.

Shipping since
2013
Senior engineer, first call to handover
1

Trusted byteams at

Services

What we build

Different things to build. The same engineer building them, from first call to production.

Web applications

Data-dense product interfaces that ship to production and survive being handed over.

  • Next.js
  • React
  • TypeScript
  • Postgres
  • Supabase

AI products

Retrieval, agents and automation against your own data, with evaluation attached.

  • RAG
  • Mastra
  • MCP
  • VectorDB

Cloud and infrastructure

Deploys, auth, monitoring, and a cost model you can predict before the invoice.

  • AWS
  • Docker
  • Terraform
  • Sentry

Marketing sites

Fast, editable sites that convert. Built to be updated by you, not re-quoted.

  • Next.js
  • Sanity
  • Vercel
  • Figma

Design systems

Tokens, components and documentation, delivered as working code you can build on.

  • Figma
  • Tailwind
  • shadcn
  • Chakra UI

Mobile and desktop

React Native and Electron builds, for when the product has to live outside the browser.

  • React Native
  • Electron

Industries

Thirteen years of other people’s hard problems

The constraints change. What makes software survive contact with real users does not.

01

Banking and finance

Regulated systems at scale

Retail banking and mortgage platforms, where an audit trail is a requirement rather than a feature.

02

Publishing and media

High-traffic editorial platforms

Content systems built for people who publish daily and cannot wait on a developer to do it.

03

Fashion and retail

Brand-grade storefronts

Commerce and campaign work where the design is the product and the tolerance for drift is zero.

04

Travel and hospitality

Booking and inventory

Search, availability and payment paths that stay correct while thousands of people use them at once.

05

Sustainability

Consumer-facing schemes

National recycling infrastructure, run as a public service with the traffic profile of a campaign.

06

Startups

Zero to production

Founder-built prototypes taken to something a paying customer and an investor can both rely on.

Process

How the work actually runs

Four steps, and the same engineer through all of them.

  1. 01

    Research and discovery

    Paid, before any number. We read the concept and what already exists: code, schema, whatever the last team left. Then an architecture, a decision log and a fixed quote. Nothing gets estimated on a hunch.

    decision · retrieval strategyversey
    • Context stuffinghits the window
    • Hybrid retrievaltwo code paths
    • RAGone code path, every model

    Written down before a line was built. No conditional logic per model.

  2. 02

    Design and build

    One step, because the same person makes the calls and writes the code. There is no handoff from a designer to an engineer here, which is why the interface and the code never drift apart and nobody has to translate between them.

    one systempodback
    • Design
    • Website
    • API
    • Database
    • Testing
    • Infrastructure
    • Email

    Seven parts of one product, and one owner across all of them.

  3. 03

    Handover as we go

    Documentation is written while the work happens, in your tools rather than ours, not assembled at the end when the reasons have been forgotten. Nothing important ends up living only in one person's head.

    every engagement ships withconfluence · notion · github
    • Architecture and decision records
    • Operational runbooks

      the routine operations, written down

    • Deployment procedures

      each release, step by step

    • Information security policies

      incident response · BCP/DR · risk register · access control

    • Onboarding and offboarding guides

    Written while the work happens. Nothing important ends up living only in one person’s head.

  4. 04

    After launch

    Shipping is the middle. Monitoring, the errors that only appear under real traffic, and the work that becomes obvious once people are actually using it. Often that turns into the next thing, and the one after that.

    client tenure20212026
    • Podbackrunning
    • Pebble5 returns
    • Versey2026

    Three retained clients at a time. The limit is what makes the response times real.

What you get

What arrives, either way

True of a six week build and a four year retainer. None of it is a line item you have to ask for.

A number you can trust

Quoted after somebody has read the system it lands in, not before.

Working software every week

Something you can open and use, rather than a status report about it.

One person accountable

The same engineer from the first call to production, making the calls and writing the code.

The reasoning, written down

A decision log with the alternatives, still readable a year later.

Tests where they matter

On the paths that cost money when they break, not a coverage number.

Deploys with no manual steps

No runbook of commands, no machine that is the only one it works from.

You hear it from us first

Monitoring and error tracking wired before launch, not after the first incident.

Everything is yours

The code, the infrastructure and the documentation, in your accounts from the first commit.

Proof

“A safe pair of hands”

Not our words. Two clients, separately, who have never met.

To say HAM Studio is a safe pair of hands is an understatement. I consider them an integral part of the team. Projects are always thoughtfully scoped, managed efficiently, and delivered on time.
Diana CaldwellHead of Marketing & Communications · Podback
You could not wish for a more capable, steady, or mature pair of hands. Julien at HAM Studio delivered ahead of schedule and impressed us with his calm approach to navigating obstacles. Any team would benefit immensely.
Adam HoggeCTO · Pebble
Strongly recommend Julien at HAM Studio if you need the tech equivalent of a one-man swat team.
Will TaylorFounder · Versey
I've worked on this project before and developers have struggled to get what we need. Julien at HAM Studio understood what we were aiming for, involved me in the build, and executed brilliantly. The ROI for the client was so good it's still paying dividends now.
Emma R.Founder · Oxford Comma Digital

Engagements

How it starts

Nothing large gets quoted before we have read the system it lands in. Whether yours needs that reading at all is what the first call settles.

  1. 01A fifteen minute callFree. Whether it is a fit, what shape the work is, whether it has an end date, and whether it needs a discovery at all.
  2. 02The Pre-Build DiscoveryA day to a week, on work large enough to warrant it. You get an architecture, a decision log and a fixed quote.
  3. 03Retainer, or a fixed scopeOngoing capacity if the work has no end. A fixed scope and a fixed number of weeks if it does.

Work with us

Three retained clients at a time

That limit is what makes the response times real, and it means there is a point at which the answer has to be no. Right now there is room.

Available now