Our engineers work inside your team until it’s in production.
Our senior engineers join you on site and on your network. They build with your engineers, work through approvals with your risk and security teams, and leave once your people can run the system.
Who joins your team
A senior team, sized to the work.
Every engagement owns an outcome with a scope and an end date, and one lead engineer is accountable for it. You don’t manage our people; you hold us to the result.
Lead engineer
Owns delivery, scope and stakeholders. Your single point of contact, and hands-on in the code.
ML engineering
Models, data pipelines and evaluation, worked through with your data owners and model validation team.
Platform engineering
Infrastructure, deployment and security controls, worked through with your infrastructure and security teams.
How an engagement runs
Working software from the first month.
weeks 1–2
Embed
Onboarding and access through your normal process. We map the workflow with the people who do it and agree how success is measured.
weeks 3–4
First working version
Running in your environment, on your data, with tests your risk team can read.
weeks 5–12
Build to production
Weekly releases and demos. Approvals worked through with risk and security as we go, not saved for the end.
handover
Your team runs it
Runbooks, training and shadowed on-call. Then we leave, or stay on a run-and-improve retainer.
Inside your security rules
We work the way your staff do.
- Background checks to your standard before anyone gets access.
- Access only through your identity and privileged-access tools, reviewed and removed when the work ends.
- Work on your managed devices or virtual desktops. No client data on Arthiq laptops or systems.
- Your change, release and incident processes, not ours.
- Contract terms for an outsourcing arrangement, including those required by MAS outsourcing guidelines and DORA Art. 30.
How it compares
Three ways to get an AI system built.
| What you get | Contractors | Consultancy | Forward-deployed |
|---|---|---|---|
| A defined outcome and end date | — | ✓ | ✓ |
| Senior engineers writing the code | ✓ | sometimes | ✓ |
| Working inside your environment | ✓ | sometimes | ✓ |
| Models, evaluation and governance in one team | — | sometimes | ✓ |
| Your team able to run it afterwards | — | sometimes | ✓ |
What you keep
Everything we build stays with you.
Code and infrastructure
In your repositories and your cloud or data centre from the first commit.
Models and tests
Weights, adapters, test sets and evaluation code.
A team that can run it
Runbooks, architecture records, and engineers who have operated the system with us.
When it ends
Exit criteria, agreed on day one.
- Your team has run the system for an agreed period with us only on call.
- Tests and monitoring are in place and owned by your people.
- Documentation is approved by your system and model owners.
- Our access has been removed and the removal confirmed.
Questions
What clients ask.
Do your engineers have to be on site?
For the first weeks and key milestones, yes. Otherwise they can work remotely through your virtual desktops, under the same controls.
Is this outsourcing under MAS or DORA rules?
Usually, yes. We provide what your outsourcing assessment needs and accept the contract terms that come with it.
Who owns what we build?
You do: code, models, tests and documentation.
What if we need you after handover?
We can stay on a run-and-improve retainer, or come back for the next system.
What we don’t promise
Some claims we won’t make.
- That we can start next week. Clearances and onboarding take time, and we plan for them.
- Open-ended engagements. Every team has an end date and exit criteria.
- Unlimited capacity. We take on as many embedded teams as we can staff with senior people, and no more.
Put engineers inside your team.
Tell us the system and the deadline. We’ll tell you who we’d send.