#164 · Primary category: MLOps & Evaluation

sematic

ai data-science machine-learning ml ml-ops ml-pipeline ml-pipelines mlops pipeline python python3

An open-source ML pipeline development platform

Project last updated:01/09/25

GitHub Stars

1.0K

Forks

63

Contributors

21

License

Apache-2.0

Why we included this project

Sematic is an open-source platform for building ML pipelines, and its main idea is that you write everything in plain Python and can run the same code on your laptop or on a Kubernetes cluster. That makes it a practical step up for teams that have outgrown ad-hoc training scripts but are not ready to commit to heavy infrastructure. It chains data processing jobs like Apache Spark together with model training in PyTorch or TensorFlow, type-checks each step at runtime, and records inputs and outputs so you can inspect a run in the web dashboard and resume from any point in the graph. Step caching and automatic retries smooth out the daily iteration loop, and the integrations with Spark, Ray, and data warehouses handle the mixed compute that real training jobs tend to need. The project grew out of experience at self-driving car companies, which shows in the attention to traceability and reproducibility.

Articles for this project

No articles for this project yet.

To suggest a topic or contribute an article, contact us.

Related projects in this category