A India-based digital marketplace connecting casting directors and producers to curated actor profiles through a cloud-native, AI-powered platform.

Impact

• Average image-validation time: < 2 s per upload
• Person-detection accuracy: 98 % (YOLOv5)
• Overall ML workflow accuracy: 90 %+ across all models
• Scales to 100 concurrent validations without latency spikes

‍

Hiring teams were drowning in spreadsheets and unverified headshots. They needed a secure, GDPR-friendly web application that would let subscription users filter thousands of actor profiles by age, location, language, and filmography—without exposing full details to non-paying visitors.

Beyond search, the business required real-time notifications, role-based access control, and Stripe-grade subscription metering so directors across North America could preview talent on the move. Finally, every uploaded image had to pass stringent validation—single face, forward orientation, no NSFW content—within two seconds to protect brand integrity.

‍

Infolitz engineered a modular Python Django backend that exposes JWT-secured REST endpoints via Django REST Framework. PostgreSQL powers lightning-fast, attribute-rich queries, while AWS S3 handles headshot storage and versioning at petabyte scale. Subscription logic ties directly into Stripe and PayPal, automatically enforcing view and connection limits and driving recurring revenue.

To amplify talent discovery, we embedded an ML image-validation pipeline. OpenCV pre-processes uploads, YOLOv5 pinpoints people, MTCNN & TensorFlow verify face orientation, and Hugging Face NSFW models block inappropriate content—all orchestrated in under two seconds. Firebase Cloud Messaging pushes match alerts, chat messages, and usage warnings instantly to iOS, Android, and web clients, keeping casting directors engaged across North America & APAC.

WebSockets enable encrypted, real-time chat once a connection request is accepted, and optional Elasticsearch indexes every profile so producers can surface niche skills—think “multilingual stunt performer”—in milliseconds.

‍

Tech Stack Highlights

• Framework: Python Django
• Database: PostgreSQL (to handle complex queries and filtering efficiently)
• Authentication: Django REST Framework (DRF) with JWT for token-based authentication
• Media Storage: Google Cloud Storage
• Search Engine: Elasticsearch (optional, for advanced filtering and sorting)
• Push Notifications: Firebase Cloud Messaging (FCM) for delivering push notifications to mobile apps
• OpenCV 4.10.0 – preprocessing
• MTCNN 1.0.0 – face detection
• TensorFlow 2.12 – accelerates MTCNN operations
• YOLOv5 – high-accuracy person detection
• NSFW JS (Hugging Face) – lightweight content moderation
• Django REST Framework 4.2.7 – API handling

Together, these gains streamline casting workflows, shorten talent shortlists, and unlock a subscription revenue engine that grows in lockstep with user engagement.

Ready to see how CastMe can revolutionize talent discovery and onboarding for your studio?

‍Book a no-obligation, 20-minute strategy call with our product architects.