What do you do?

Aporia is a monitoring platform for machine learning models in production. With Aporia’s monitor builder, data science teams can easily create monitors for detecting drift, unexpected bias and integrity issues, and receive live alerts to enable further investigation and root cause analysis.

With on-VPC deployment and seamless integration with any ML infrastructure and workflow, Aporia makes ML monitoring simple and secure.

Using Aporia, data scientists and ML engineers have a live view of their models in production in one place – with real-time tracking for model activity, inference trends, data behavior, and actual model performance.

How much does it cost?

Aporia is free to use for up to 1 GB with Aporia’s cloud deployment.

Professional plans include on-VPC deployment and unlimited models. Pricing is based on the number of predictions and data retention.

Learn more about Aporia’s pricing.


What’s a sample use case? Where can I learn from?

  • Fraud Detection
  • Recommender System
  • Churn Prediction
  • Lead Prioritization
  • Sales Forecast

Feature List

  • Visibility
    • Watch live inferences
    • Explore actual performance (F1, Accuracy, Precision, …)
    • Analyze behavior of different data segments
  • Concept Drift Monitoring
    • Define monitors for drifting features
    • Detect when predictions are drifting
    • Assure models retrain occasionally
    • Define custom performance metrics
  • Data Integrity Monitoring
    • Detect change in missing values rate
    • Identify unexpected values
    • Monitor model’s activity and detect outages
    • Detect out of schema features
  • Root-Cause Analysis
    • Live productions distributions
    • Explore data statistics
    • Slice & Dice in production data
  • Setup & Integrations
    • Deployed on your VPC of choice (AWS, GCP, Azure)
    • SDK Integration
    • Amazon S3 / Azure Blob Storage / Google Cloud Storage integrations
    • REST API
    • Get alerts to Slack / MS Teams


There are no reviews yet. Be the first to write one.