Candidate data at volume,
on a machine you control.
Private LLM hosting on a dedicated machine in EU-Central, for staffing and temp agencies screening, parsing, and matching high volumes of candidate resumes and personal data.
Not legal advice. This page describes GPUwerk's infrastructure and contractual commitments. Whether a given use of candidate data has a valid legal basis under GDPR or other applicable law is a decision for your agency's own counsel.
Why staffing agencies ask about this
A staffing agency's core asset is a database of candidate resumes, contact details, and work history, often at a scale most other businesses don't handle. Running that data through a public AI tool for screening or matching sends personal data on thousands of people to infrastructure the agency doesn't control.
"We screen thousands of resumes and don't know where they end up"
The hardware behind a GPUwerk instance is physically in EU-Central, operated by PRINT IT! SE, a Societas Europaea registered in Prague. Under the data processing agreement, there are no sub-processors for instance content, and GPUwerk does not access your workspace's content except at your request for support or where a legal obligation requires it.
"Billing per document doesn't work at our volume"
GPUwerk bills by instance-hour, not per document or per candidate processed. A resume-screening or matching workload that runs continuously through the day costs the same whether it processes a hundred records or ten thousand.
"Recruiters already paste resumes into consumer AI tools"
Staff summarizing a resume or drafting a candidate outreach message in a public tool is a real exposure, not a hypothetical one, particularly with candidate contact details in the prompt. A sanctioned alternative on your own instance, same chat interface, keeps that workflow off third-party infrastructure. See the private ChatGPT setup →
What this is, and what it isn't
| Question | Answer |
|---|---|
| Is this an applicant tracking system? | No. GPUwerk hosts general-purpose LLM infrastructure. You run your own parsing, screening, or matching software on the instance. |
| Does pricing scale with candidate volume? | No. Billing is by instance-hour, not by record processed. |
| Where is data physically processed? | EU-Central, on a dedicated single-tenant machine assigned to you |
| Who can access instance content? | Through the instance itself, only holders of your SSH keys; password login is disabled fleet-wide. GPUwerk keeps infrastructure administrator access, as on any hosted service, and does not use it on your content except at your request for support or where a legal obligation requires it. |
| Sub-processors for the workload? | None, listed at /legal/sub-processors |
| DPA (GDPR Art. 28)? | Published at /legal/dpa, no charge |
| Certifications held? | None claimed on this page or elsewhere on the site. Verify directly with us before relying on this page if your procurement requires one. |
Pricing
On-demand billing starts at $0.79/hour for a spark-1x instance and $1.79/hour for spark-2x. A held (reserved) allocation runs at 75% of the on-demand rate: $0.59/hour and $1.34/hour respectively, worth considering for an agency with steady day-to-day screening volume rather than occasional spikes. Full details on the pricing page.
Questions we get from staffing agencies
We process thousands of resumes a month. Does the pricing model work at that volume?
Yes. Billing is by instance-hour, not by document or candidate processed, so resume screening, parsing, or matching at high volume doesn't add per-record cost. On-demand starts at $0.79/hour for spark-1x; a held allocation drops that to $0.59/hour if usage is steady enough to reserve capacity.
Does this meet GDPR requirements for candidate data?
GPUwerk provides infrastructure evidence relevant to that assessment: EU-Central hardware and operations, a dedicated single-tenant machine, an Article 28 DPA, and no sub-processors for instance content. Whether a specific use of candidate data has a valid legal basis under GDPR is a decision for your agency as controller.
Where is candidate data physically processed?
EU-Central, on a dedicated single-tenant NVIDIA DGX Spark machine assigned to your account, operated by PRINT IT! SE, a Societas Europaea registered in Prague, Czech Republic.
Do you sign a DPA covering candidate and client data?
Yes, a standard GDPR Article 28 DPA is published at /legal/dpa at no charge, and there are no sub-processors for instance workloads.
Related pages: HR and recruiting agencies, consulting firms, and the private LLM hosting overview.
Related pages
Bring your candidate-data workload to a machine you control.
Talk to the people who run the racks, or start with a pilot and a practical rollout plan.
Talk to us Deploy an instance