Home AI data stack MLOps and feature stores

MLOps and feature stores

Training, deployment and monitoring of machine-learning models.

Showing 14 tools
Loading tools…

Kedro

Pipeline framework

Open-source Python framework for structuring data and ML pipelines: project template, Data Catalog, nodes, Kedro-Viz. Free.

Visit site

Metaflow (Outerbounds)

Pipeline framework

Netflix's open-source Python framework for ML pipelines: automatic artifact versioning, resume, laptop-to-cloud scaling. Managed tier via Outerbounds.

Visit site

ZenML

Pipeline framework

Open-source Python framework for portable ML pipelines: one codebase runs on 14 orchestrators, with a hosted Pro control plane from $999 a month.

Visit site

Kubeflow

Pipeline framework

Open-source ML platform for Kubernetes: Python pipelines (KFP) plus distributed training, tuning, serving and a model registry, CNCF graduated.

Visit site
Free OSS; infrastructure extra

Flyte

Open-source orchestrator

Open-source workflow runtime for data, machine learning, and AI workloads, with Python task authoring, recovery, caching, and resource configuration.

Visit site
Free open-source core

DVC / DataChain

Experiment tracking

DVC versions data and pipelines alongside Git. DataChain separately processes and curates datasets. DVC is now maintained by lakeFS.

Visit site

Arthur AI

Monitoring

AI observability platform that monitors ML model drift, traces and evaluates LLM agents, and governs agents, with a free tier and MIT Engine.

Visit site

TrueFoundry

ML platform

Kubernetes-native platform for deploying models, jobs and LLMs in your own cloud, plus an AI Gateway for routing, budgets and guardrails.

Visit site

JFrog ML

ML platform

End-to-end MLOps platform (formerly Qwak) for building, serving and monitoring models inside the JFrog Platform, with Artifactory as registry.

Visit site

Chalk

Feature store

Real-time feature platform: define ML features in Python, compute them at request time, and build point-in-time training sets.

Visit site

ClearML

ML platform

End-to-end open-source MLOps: experiment tracking, GPU orchestration with fractional GPUs, dataset versioning, and pipelines at $15/user.

Visit site

Comet

Experiment tracking

Experiment tracking with bundled production drift monitoring, plus Opik, its Apache-2.0 open-source LLM evaluation product.

Visit site
12