ZenML Labs

ZenML offers flexible deployment options for AI workflows, including SaaS, cloud, and self-hosted setups.

ZenML automates AI workflows, eliminating manual deployments and non-reproducible experiments with production-ready pipelines.

The page provides instructions for building the ZenML brick set, available as a downloadable PDF.

Teams use ZenML to integrate AI platforms, supporting batch evaluations, real-time serving, and GenAI workflows.

Blog posts by Alex Strick van Linschoten on the ZenML blog.

Blog posts by Alexej Penner on the ZenML blog.

The guide explains architecture patterns, best practices, and strategies for scaling machine learning infrastructure.

How models behave inside real agents, measured as cost per correct answer, with the method, noise and caveats published for every experiment.

ZenML is hiring to expand its open-source team.

One plan, $39 a month: your AI engineer on one AI agent, one ML project and your whole team. Open source is free; more agents, projects and on-prem are a conversation.

The page assesses ZenML’s AI orchestration platform for company use.

The page compares eight alternatives to Kubeflow, addressing challenges like complexity and operational overhead, including ZenML’s developer-friendly approach.

Side-by-side comparisons of ZenML (AI orchestration) and Kitaru (agent replay) against other tools and platforms.

Arize AX and Phoenix trace, score, and identify issues. Kitaru imports Phoenix traces and re-executes agents using recorded tool calls.

Braintrust scores tasks against datasets. Kitaru re-executes served sessions using recorded tool calls and includes a Braintrust importer.

Blog posts by Adam Probst on the ZenML blog.

ZenML has worked with leading companies to deploy AI in production for five years. The team behind ZenML and Kitaru includes an AI engineer.

The ZenML team’s AI engineer and Kitaru analyze traces alongside evaluations to identify patterns and create checks.

How to reach the ZenML team: email hello@zenml.io, join the community Slack, or book a demo.

The AI engineer compares agent traces to existing evaluations, identifies gaps, and converts findings into verifiable checks. It can fine-tune open models using the same traces and integrates with open-source Kitaru and ZenML.

Legal company information for ZenML GmbH, Munich, Germany.