Technical overview of how TalentScreen is built
TalentScreen is built on modern, scalable infrastructure designed for performance, security, and reliability. This article provides a high-level overview of the platform's technical foundation.
TalentScreen uses a serverless architecture deployed on Cloudflare Workers. This ensures low latency globally and automatic scaling during high-traffic periods. Authentication relies on magic links instead of passwords, reducing security risks and improving user experience.
Data residency
All candidate data is stored in Supabase's secure PostgreSQL database with encryption at rest and in transit.
Session initialization
When a candidate starts, a session record is created in the database with a unique ID and timestamp.
Auto-save mechanism
Every question response triggers an API call saving answers to the database. If connection drops, the last saved state is recoverable.
AI processing
For AI interviews, audio is transcribed in real-time and sent to OpenAI for response generation. Responses are scored based on predefined rubrics.
Report generation
On submission, the platform aggregates scores, computes analytics, and generates a PDF report stored in Cloudflare R2.
TalentScreen implements industry-standard security practices including role-based access control, encrypted data storage, regular security audits, and GDPR-compliant data handling. Candidates can request data deletion at any time through the privacy settings.
Was this article helpful?