How we handle your systems and data
This page is written for the person on your side who has to ask. It describes what we do with the access you give us, where your data goes, and what happens when the project ends. There is no marketing here, just the practice.
- Named accounts with least access
- Credentials only in a password vault
- Your data stays in your environment
- Access removed and confirmed at handover
Access
Each engineer on your project gets their own named account in your systems, with the least access the work needs. We do not share logins between people. Admin rights are requested only when a task needs them and given back when it is done. Two-factor authentication is switched on wherever your system supports it.
Passwords and keys
Credentials you give us are stored in a password vault, never in chat messages, email or documents. Where you have your own vault, we ask you to share through it. API keys are created with the narrowest scope available and rotated at handover.
Your data
We work inside your environment, not on exported copies. Where test data is needed, we use scrubbed or generated records rather than real customer details. Backups we take during a migration are stored in your accounts and deleted once you confirm the migration is complete.
Code and ownership
Code lives in your repository from the first commit. You can see every change as it happens. The intellectual property in what we build for you is yours on final payment, and we do not reuse your code for other clients.
Devices and office
Our team works on company-managed laptops with disk encryption, screen lock and up-to-date systems. Client work is not done on personal devices. Office access is controlled and visitors are accompanied.
AI data and models
For private AI projects, the models, documents, embeddings and outputs stay inside your server or your cloud account. Nothing is sent to a third-party AI provider unless you choose that route, and then only under your own account, with data retention turned off, on data you have approved. We do not train models on one client’s data for another.
If something goes wrong
You have a named contact for security questions. If we discover an incident affecting your systems or data, we tell you as soon as we know, with what happened, what we have done and what we recommend. You decide what is communicated to your customers.
What we do not have yet
We are not SOC 2 or ISO 27001 certified. We follow the practices above because they are right, and we can walk your IT lead through them on a call. If your procurement process requires a certification, tell us early and we will discuss what is possible.
Bring your IT lead to the call
We are happy to answer security questions before any scope is agreed.