it is a great platform for data scientists who want to create ml pipelines and build those pipelines. there is no complexity to creating those pipelines. Review collected by and hosted on G2.com.
it is not much reliable and also employee faces lot of complexity to congure it Review collected by and hosted on G2.com.
Kubeflow helps us in addressing requirements for each stage in the ML lifecycle, from exploration through to training and deployment, we use Kubeflow for building the ML pipelines most, it is fast compared with Apache Airflow Review collected by and hosted on G2.com.
we used to use Airflow earlier, we faced little difficulty in setting up Kubeflow due to limited documentation, once it was done, we are comfortable in using it. Review collected by and hosted on G2.com.
Very user friendly and easy to use also making my work life easy Review collected by and hosted on G2.com.
Sometimes i use to face slowness issues but it's manageable and not an big issue Review collected by and hosted on G2.com.
Ability to seamlessly experiment with diffreent parameters and store results. Review collected by and hosted on G2.com.
The integration with Python notebooks is a bit tricky with not much clear guidelines.
Lack of proper documentaion Review collected by and hosted on G2.com.
Possibility to handle all ML model lifecycle phases in the same place Review collected by and hosted on G2.com.
Sometimes documentation is difficult to follow in on-prem scenario Review collected by and hosted on G2.com.
It provides an oppertunity to make all the things happen on cloud and makes the Data Scientist/ML engineer responsibilities easy. Review collected by and hosted on G2.com.
Nothing as of now. As far as I have worked on Kubeflow ... Review collected by and hosted on G2.com.
Pipelines in Kubeflow made my orchestration process very easy. Review collected by and hosted on G2.com.
Management of artifacts. May be a direct integration with S3 and other storages through UI Review collected by and hosted on G2.com.