Case Study: Visii achieves 3.5x more cost-effective infrastructure and 5x greater agility with Amazon Web Services

How Visii builds its visual search technology using AWS for machine learning

Visii is a London-based startup that uses machine-learning–driven visual search to improve product discovery on client websites. As the business prepared to scale, its original managed-cloud setup proved inflexible and couldn’t reliably handle spikes in image-search traffic; the in-house messaging system risked jamming ML pipelines, DevOps was a single-person bottleneck, and upfront reserved compute costs limited agility.

Visii migrated its stack to AWS, adopting Amazon SQS to buffer image requests, Amazon EC2 (including Spot Instances), RDS, Elastic Load Balancing, Auto Scaling, and Route 53. The move made infrastructure 3.5× more cost-effective, cut monthly compute costs by 45% using Spot Instances, increased deployment agility fivefold, removed a single point of failure, and gave the team startup support through the AWS Activate program.

Visii

George Whitelaw

Chief Technology Officer