What challenges did Sonrai face before AI implementation?

Datasets contained thousands of potential biomarkers but only hundreds of patient samples. The team needed to evaluate hundreds of modality and model combinations without manual experiment tracking. Clinical use required full traceability from raw data through model decisions for regulatory submissions.

What was the impact of the AI solution?

The end-to-end pipeline executed in under 10 minutes. Sonrai modeled 8,916 biomarkers and performed hundreds of experiments with full lineage. Time spent curating data for biomarker reports was reduced by 50%.

Source

Evidence: Low35/100

Sonrai uses Amazon SageMaker AI to accelerate precision medicine trials

Use case typePrecision medicineUpdated Jun 13, 2026

Sonrai built an end-to-end MLOps framework on Amazon SageMaker AI for precision medicine biomarker discovery. The workflow helps evaluate many omic combinations while preserving traceability and reproducibility required for regulated clinical use.

Organization: Sonrai
Industry: Healthcare
Location: Ireland
Published: February 2026

Reported outcomes

−50%

time spent curating data for biomarker reportsTime & speed

10 minutespipeline runtime8,916 countbiomarkers modeled+94%sensitivity of top model+89%specificity of top model

Strategic outcomes

New product / capabilityBuilt end-to-end biomarker discovery pipelineRisk & compliancePreserved traceability for regulated clinical useSpeed & agilityEnabled rapid end-to-end pipeline executionBetter decisions & insightSupported large-scale experiment tracking

Catalog median for time & speed deployments: −50% across 312 reported metrics. Compare benchmarks →

Why do we believe this?Outcome claims, sources, and evidence checks

Normalized claim

Pipeline runtime: 10 minutes decrease

AWS Machine Learning BlogFeb 23, 2026Blog postExplicit claimLow evidence strength

“the entire pipeline—from raw data to final models and reports—executed in under 10 minutes”

Normalized claim

Biomarkers modeled: 8,916 count increase

AWS Machine Learning BlogFeb 23, 2026Blog postExplicit claimLow evidence strength

“8,916 biomarkers modeled and tracked”

Normalized claim

Time spent curating data for biomarker reports: 50% decrease

AWS Machine Learning BlogFeb 23, 2026Blog postExplicit claimLow evidence strength

“50% reduction in time spent curating data for biomarker reports”

Normalized claim

Sensitivity of top model: 94% increase

AWS Machine Learning BlogFeb 23, 2026Blog postExplicit claimLow evidence strength

“achieving 94% sensitivity”

Normalized claim

Specificity of top model: 89% increase

AWS Machine Learning BlogFeb 23, 2026Blog postExplicit claimLow evidence strength

“89% specificity with an AUC-ROC of 0.93”

Why do we believe this deployment?Customer identity, provider attribution, maturity, and source checks

Customer: Sonrai
Provider: AWS
Maturity: Unknown
Linked source: AWS Machine Learning Blog

No explicit deployment-stage evidence found.

Customer identity supportedSource describes one deploymentMaturity evidence evaluated

Primary read

Use case focus

Showing 3 of 3

1MLOps for life sciences
2Precision medicine analytics
3Biomarker discovery automation

Datasets contained thousands of potential biomarkers but only hundreds of patient samples.
The team needed to evaluate hundreds of modality and model combinations without manual experiment tracking.
Clinical use required full traceability from raw data through model decisions for regulatory submissions.

Sonrai used Amazon S3 for secure data repositories and Amazon SageMaker AI for development, training and deployment.
The workflow used SageMaker Studio, Code Editor, JupyterLab, MLflow experiment tracking, SageMaker Pipelines and the SageMaker Model Registry.
The pipeline generated reports with Quarto and supported model promotion, validation and deployment options for inference or batch validation.

Technologies

Amazon SageMaker AI Amazon S3 Amazon SageMaker Studio Amazon SageMaker model registry Amazon SageMaker Pipelines MLflow

The end-to-end pipeline executed in under 10 minutes.
Sonrai modeled 8,916 biomarkers and performed hundreds of experiments with full lineage.
Time spent curating data for biomarker reports was reduced by 50%.

Architecture

End-to-end MLOps framework on Amazon SageMaker AI with Amazon S3 data repositories, SageMaker Studio, Code Editor, JupyterLab, MLflow experiment tracking, SageMaker Pipelines, SageMaker Model Registry, Quarto report generation and deployment to SageMaker endpoints or batch validation.

Sources & evidence1

Evidence: Low35/100Evidence strength