What is the ML Monitoring skill for Claude Code?

It is a specialized capability that provides patterns and best practices for implementing model observability, drift detection, and performance tracking within Databricks environments.

Can I use this with Databricks Lakehouse Monitoring?

Yes, the skill includes specific code patterns for programmatically creating and managing quality monitors via the Databricks Workspace Client.

How does this skill detect model drift?

The skill utilizes statistical methods, such as the Kolmogorov-Smirnov test, to compare reference data against current production data to identify distribution shifts.

Does it support alerting for production issues?

Yes, it provides implementation patterns for monitoring inference tables and triggering alerts when latency or error thresholds are exceeded.

ML Monitoring for Databricks

Name: ML Monitoring for Databricks
Author: vivekgana

byvivekgana

•

数据科学与机器学习

Implements automated model monitoring, drift detection, and performance tracking for production machine learning systems on Databricks.

关于

This skill empowers AI developers to build robust MLOps pipelines by integrating comprehensive monitoring into their Databricks workflows. It provides standardized patterns for setting up Databricks Lakehouse Monitors, performing statistical drift detection using KS tests, and tracking real-time inference performance metrics. By automating these critical observability tasks, users can ensure their production models remain accurate, reliable, and performance-aligned with business requirements through standardized logging and alerting patterns.

主要功能

Statistical feature drift detection using Kolmogorov-Smirnov tests
Standardized patterns for inference log capturing and analysis
2 GitHub stars
Real-time performance tracking for inference latency and error rates
Automated Lakehouse Monitor configuration via Databricks SDK
Custom alerting logic for production ML system health

使用场景

Monitoring inference latency to ensure SLA compliance in real-time applications
Detecting feature drift in production models to trigger retraining workflows
Building automated quality dashboards for enterprise ML model governance

关于

主要功能

Statistical feature drift detection using Kolmogorov-Smirnov tests
Standardized patterns for inference log capturing and analysis
2 GitHub stars
Real-time performance tracking for inference latency and error rates
Automated Lakehouse Monitor configuration via Databricks SDK
Custom alerting logic for production ML system health

使用场景

Monitoring inference latency to ensure SLA compliance in real-time applications
Detecting feature drift in production models to trigger retraining workflows
Building automated quality dashboards for enterprise ML model governance