AI and machine learning

Embeddings and vector search

We turn text and images into embeddings and search them at scale, alongside classic keyword search and perceptual image hashing.

What we deliver

  • Image embeddings with CLIP, computed in batch on GPU or on demand in serverless functions
  • Vector indexes in OpenSearch next to the keyword index
  • Perceptual hashes and colour histograms for near-duplicate images
  • Candidate generation with blocking, so millions of comparisons become thousands

Typical uses

  • Matching tens of thousands of unmatched products against a catalogue of 200,000+
  • Finding the same product photographed by different shops
  • Semantic search over documents and support content

Have something in mind?

Tell us what you want to achieve. We reply with how we would approach it and what it would take.

Start a project