Skip to content

Trust center / Security

Reviewed 20 October 2026

Protection at every boundary.

Company access, candidate work and private assessment evidence have different permissions. Our security design keeps those responsibilities separate.

Verified company access

Work email, email verification and manual approval before workspace access.

Isolated execution

Candidate code runs with bounded resources and question-scoped access.

Private evidence

Recordings and submissions require authorized access.

Company identity and access

Company registration requires a work email, email verification and manual review before access is approved. Bot protection and server-side validation protect the access flow. Approval confirms platform access; it is not an endorsement or independent certification of the company.

Server and database checks enforce organization membership and permitted operations. A company name, URL, browser identifier or disabled button is not an authorization decision. Workspace owners should invite only authorized colleagues, assign appropriate roles and remove access when it is no longer needed.

Candidate access and assessment integrity

Private invitation links have an expiry and are exchanged for an attempt-bound session. Invitation secrets are stored as hashes. Time limits, assigned questions, AI permissions and submission state are enforced by the service, rather than trusted to browser state alone.

The interview version fixes the assessment configuration for its cohort. Changes requiring different questions, timing or AI conditions need a new version. Hidden graders, answer keys and reference solutions are separated from candidate content and are not supplied to the candidate assistant.

Code execution and AI boundaries

  • Candidate execution is authorized for the assigned attempt and question, with limits on runtime, resources, output and network access.
  • Grading runs separately from the candidate workspace. Private expected results are not part of the candidate package.
  • AI credentials remain on the server. Model access is controlled by the approved interview configuration, phase and allowance.
  • Provider requests and usage are recorded for accounting and investigation. A model shown in a public estimate is not automatically enabled for a real assessment.

These controls reduce the scope of a failure; they do not make arbitrary code, external models or generated outputs infallible. The provider list describes the services that can participate.

Storage, encryption and evidence access

The production design uses encrypted connections, encrypted storage and private object access. AWS evidence and recording storage is configured for encryption with managed key controls. Access to objects is scoped and time-limited; possession of a filename alone does not authorize retrieval.

Service identities have distinct responsibilities for execution, evaluation, evidence and deletion. Secrets are excluded from browser bundles and candidate sandboxes. Operational logging is designed to avoid unnecessary candidate content or credentials.

Retention settings, legal holds and deletion records are covered by the privacy notice. We do not claim customer-held encryption keys, end-to-end encryption that prevents all service access, or guaranteed recovery of deleted evidence.

Development, operations and incidents

Our release process includes automated tests, type and build checks, dependency and secret checks, and checks that confidential authoring content is excluded from public builds. Production capabilities also require their deployment-specific verification before activation.

Our incident process covers investigation, containment, preservation of relevant evidence, recovery and communication. Where a personal-data breach affects a customer, notification and assistance follow applicable law and the customer agreement.

Backups, restoration, monitoring and response arrangements must match the deployed service. Any uptime, recovery or support commitment must be stated in the applicable agreement.

Assurance and customer review

Contact us about your security requirements before sharing sensitive data or inviting candidates. We can discuss the applicable architecture, provider scope and contractual requirements. SSO/SAML, dedicated hosting, fixed data residency and a particular service level are not included unless expressly agreed and available.

Report a vulnerability

Send the affected URL, a concise description, safe reproduction steps and potential impact. Use your own account and synthetic data. Stop testing if you encounter another person’s information. Do not download it, expose private assessment content, disrupt interviews or continue exploiting the issue.

We will arrange a secure way to exchange sensitive details if needed. This reporting channel is not authorization for intrusive testing and does not promise a bounty or a particular response time.

Contact AIDriven Technologies Private Limited, the operator of BuildInterview in Delhi, India, at hello@buildinterview.com for security or responsible-disclosure questions. Include the relevant page or hiring company and a short description. Do not email passwords, invitation links, identity documents, medical records or candidate recordings. We will arrange an appropriate channel if more information is needed.

BuildInterview is operated by AIDriven Technologies Private Limited.

Back to top ↑