Where data science teams
monitor their models

Experience the new standard for post-deployment data science.
Without labels. Research Driven.

Explore the NannyML way

Crystal clear monitoring like never before.
Built around a Performance-Centric workflow, NannyML Cloud lets you monitor what truly matters, find what is broken and fix any issue.

1

Monitor what matters

Focus on one single metric to measure how your model is doing.
Get alerted when it drops.

swap_driving_apps_wheel

Quantify the impact of data drift on model performance without target data

Know the performance of your ML models 24/7. NannyML estimates the performance of your ML models taking into account historical model predictions and the current distribution of your data. All of this is possible even when ground truth is delayed or absent.

attach_money

Define a cost-benefit matrix and measure the business impact of your models

Tie the performance of your model to monetary or business oriented outcomes. So that you always know what your ML brings to the table.

warning

Avoid alert fatigue

Traditional ML monitoring look for data drift which tends to overwhelm teams with many false alarms, because not all data drifts impact model performance.
By focusing on what matters NannyML alerts are always meaningful.

Configure once. Monitor all the time.

Built on top Open Source nannyML.
NannyML Cloud solves all the infrastructure nuances. And brings extra capabilities to the table.

swap_driving_apps_wheel

Effortless observability

Know the performance of your ML models 24/7. NannyML estimates the performance of your ML models. Even if the ground truth is delayed or absent.

attach_money

Concept drift detection

Tie the performance of your model to monetary or business oriented outcomes. So that you always know what your ML brings to the table.

warning

Deployed in your cloud. Safe and Secure.

Traditional ML monitoring tend to overwhelm teams with many false alarms. By focusing on what matters NannyML alerts are always meaningful.

2

Find what is broken

Understand the root cause of the performance drop in a matter of minutes.
Use advanced tools at your disposal.

bolt

Detect Concept Drift impact on model performace

Measure the impact of concept drift on your model's performance. The presence of concept drift is the best trigger for model retraining, even better than realized performance alerts.

analytics

Uncover the most subtle changes in the data structure

• Find both global and local changes in data structure using multivariate drift detection.
Go on a granular investigation with univariate drift detection methods to determine which feature alerts are correlated with performance issues.
• Perform continuous data quality checks to ensure that the model inputs and outputs maintain high quality standards.

lightbulb

Intelligent alert ranking

NannyML links data drift alerts with the performance changes. So you can easily detect which features are causing the performances issues.

3

Fix it

No more sleepless nights without knowing how your models
are doing.

nannyml_webhook_response.json
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
{
  "header": {
    "model_name": "Car Loan Prediction",
    "model_id": 247,
  },
  "payload": {
  
"performance_alerts": { "cbpe": [ { "metric_name": "f1" } ] },
"other_alerts": { "univariate_drift": [ { "metric_name": "car_value" } ] } } }
start

Trigger retraning actions by Webhooks

Trigger a model retraining process when concept drift is detected, your estimated performance is degrading or when a heuristic rule between performance and covariate drift is met.

sdk

Automate monitoring data collection with the SDK

Leverage NannyML Cloud SDK to automate monitoring data ingestion.

tactic

Design the best issue resolution strategy

Use all the insights learned from performance estimation, concept drift, and covariate shift results to implement the best resolution technique to your model performance issue.

Configure once. Monitor all the time.

Built on top Open Source nannyML.
NannyML Cloud solves all the infrastructure nuances. And brings extra capabilities to the table.

visibility

Effortless observability

Start monitoring in minutes. Provide your model information, a reference and analysis set to understand how you model is doing.

speed

Concept drift detection

Get access to nannyML's most powerful algorithm yet. Measure the impact of concept drift on your model's performance.

CLOUD

Deploy in your cloud

Safe and Secure. Your data will never leave your cloud.

sdk

Automated monitoring

Automated infrastructure for alerting, scheduling and data collection. Leverage the SDK to do the repetitive tasks so you can focus on what matters.

Try out for free with your cloud provider
What Data Science Leaders have to say about NannyML
NannyML Logo
The Open Source library for post deployment data science