Scaling AI Visual Search

We developed a cloud-native, event-driven platform to scale Canto’s AI Visual Search to thousands of customers. Our solution supports large-scale processing of customer assets and enables efficient real-time search across millions of assets. We achieved this by implementing distributed inference for asset encoding and leveraging distributed vector databases for scalable, low-latency retrieval.

The Canto AI Visual Search solution is built upon our work on the Merlin Accelerated Intelligence Suite, leveraging its machine learning models and search capabilities.

We also created research prototypes for Canto, most notably a hybrid search prototype which extends the capabilities of AI visual search to incorporate customer-specific asset metadata.

References:

• Youtube