What challenges did Encord face before AI implementation?

Customers building physical AI and healthcare imaging solutions faced a data bottleneck in organizing and curating huge multimodal datasets. Manually finding edge cases in petabytes of raw video footage was time-consuming and slowed model iteration.

What was the impact of the AI solution?

Customers reduced their total dataset size by 35%. Customers achieved 60% faster model training and evaluation. Curated datasets led to over 20% improvement in model performance. Encord's Google Cloud usage grew by 350% in one year.

Source

GCP

Encord curates data with Google Cloud to accelerate AI deployment in healthcare, robotics, and autonomous driving

Use case typeAI model trainingUpdated 7 hours ago

Encord provides a data layer for AI development that turns messy multimodal information such as video footage and medical scans into high-quality training data. The company migrated to Google Cloud to build a platform that continuously processes data, identifies critical events, and helps customers curate and annotate training data faster.

Organization: Encord
Industry: Tech & Comms
Location: United Kingdom
Published: July 2026

Reported outcomes

350%

Google Cloud usage growthRevenue & growth

−35%customer dataset size reduction−60%model training and evaluation speed+20%model performance improvement

Strategic outcomes

Scale & capacityScaled to petabyte-level multimodal data processingCost efficiencyStreamlined radiology and research annotation workflowsSpeed & agilityAccelerated AI model iteration for robotics and medical AI teams

Catalog median for revenue & growth deployments: +35% across 152 reported metrics. Compare benchmarks →

Primary read

Use case focus

Showing 3 of 3

1AI model training
2Data platform modernization
3Medical imaging

Customers building physical AI and healthcare imaging solutions faced a data bottleneck in organizing and curating huge multimodal datasets.
Manually finding edge cases in petabytes of raw video footage was time-consuming and slowed model iteration.

Encord built its platform on Google Cloud to continuously process data and automatically select and label relevant footage.
The company uses Cloud GPUs on Compute Engine and Cloud SQL, and it built a DICOM application with Google Cloud Medical Imaging Suite to streamline radiology and research workflows.

Technologies

Google Cloud Cloud GPUs Compute Engine Cloud SQL Google Cloud Medical Imaging Suite Google Cloud Marketplace Google Kubernetes Engine Gemini

Customers reduced their total dataset size by 35%.
Customers achieved 60% faster model training and evaluation.
Curated datasets led to over 20% improvement in model performance.
Encord's Google Cloud usage grew by 350% in one year.

Architecture

Encord runs its data-curation platform on Google Cloud using Cloud GPUs on Compute Engine for processing, Cloud SQL for infrastructure, and a DICOM application built with Google Cloud Medical Imaging Suite to connect radiology and research workflows with Encord annotation capabilities.

Sources & evidence1

Groundedness: 5/5Type: Customer StoryPublished: Jul 18, 2026Publisher: Google CloudEvidence: PrimaryConfidence: High

Primary source

AI-generated summary. Verify important details with the linked sources before relying on this case.

Explore related AI use cases

Customer evidence