Case Study: Nyris achieves efficient, multi-cloud continuous learning and faster model training with Valohai

A high-performance visual search engine

Nyris is building a high-performance visual search engine that understands image content and powers use cases across retail, manufacturing and automotive. Their challenge was running continuous experiments and custom models at scale while keeping costs and DevOps overhead low, managing datasets and versions, and avoiding cloud lock‑in. To address this they adopted Valohai’s orchestration and experiment/data‑versioning platform for their training pipelines.

Valohai enabled Nyris to allocate training resources across AWS, GCP and Azure, automate provisioning and teardown of VMs, and version experiments and datasets with a full audit trail. The result is a continuously learning pipeline with more efficient use of compute, fewer DevOps tasks (freeing engineers to improve availability and deployments), and faster iteration for data scientists focused on model research and retraining.

Nyris

Markus Lukasson

Chief Technology Officer