polyaxon/polyaxon

Health v2 · maintenance only

Flags: no_releases

Adoption not part of the score

3723 stars · 330 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-29, confidence not recorded

Polyaxon is an open-source AI engineering control plane that teams self-host on their own Kubernetes clusters to schedule, track, and govern machine learning workloads such as training jobs, sweeps, DAG pipelines, notebooks, and model serving. It provides experiment tracking, hyperparameter optimization, queue- and preset-based scheduling, RBAC governance, and a CLI/SDK/API client surface, with managed cloud offerings available.

Use cases

• schedule and track ML training experiments on Kubernetes
• run hyperparameter sweeps and distributed training at scale
• self-host an MLOps platform with experiment tracking and model registry
• turn GPU servers into shared self-service resources for a data science team
• orchestrate DAG pipelines and batch jobs for machine learning
• deploy models and notebooks as services with queues and presets
• track metrics and artifacts from PyTorch, TensorFlow, and Keras training
• govern ML workloads with RBAC, audit logs, and team-level access control

When to choose

• you need a self-hosted, Kubernetes-native MLOps control plane with full data sovereignty
• your team runs many training jobs and needs scheduling, queues, GPU sharing, and experiment tracking in one place
• you want reproducible ML workflows with Polyaxonfile specs, caching, and lineage
• you need hyperparameter optimization, sweeps, and DAG orchestration integrated with tracking

When to avoid

• you only need lightweight local experiment tracking without cluster orchestration
• you don't run Kubernetes or can't operate a self-hosted platform
• you want a fully managed turnkey SaaS with zero infrastructure setup
• your use case is simple CI/CD without ML-specific workload management

Facets

service · maturity active

workflow-automation scheduling monitoring machine-learning llm-training container-orchestration deployment api-framework cli sdk machine-learning deep-learning data-science cloud-computing self-hosted python go self-hosted cloud cli mlops experiment-tracking hyperparameter-tuning ai-control-plane gpu-scheduling model-serving notebooks distributed-training agent-sandboxes helm devops kubernetes docker web-server

Member repositories

For agents

markdown · JSON · MCP: product_card(name="polyaxon/polyaxon")