Kubeflow, MLFlow and beyond - augmenting ML delivery

Kubeflow, MLFlow and beyond - augmenting ML delivery

Presented By Stepan Pushkarev
Stepan Pushkarev
Stepan Pushkarev
CTO at Provectus

Stepan Pushkarev is a CTO of Provectus. His background is in the engineering of data platforms. He spent the last couple of years building continuous delivery and monitoring tools for machine learning applications as well as designing streaming data platforms. He works closely with data scientists to make them productive and successful in their daily operations.

Presentation Description

Reproducible ML pipelines in research and production with monitoring insights from live inference clusters could enable and accelerate the delivery of AI solutions for enterprises. There is a growing ecosystem of tools that augment researchers and machine learning engineers in their day to day operations. Still, there are big gaps in the machine learning workflow when it comes to training dataset versioning, training performance and metadata tracking, integration testing, inferencing quality monitoring, bias detection, concept drift detection and other aspects that prevent the adoption of AI in organizations of all sizes.
In this webinar, we’ll design a reference machine learning workflow. We’ll review open source tools that contribute to this workflow and are applicable to build reproducible automation of it.


Takeaways:
– A deeper view on traps and pitfalls on each stage of ML lifecycle.
– Reference implementation and automation of ML Workflow.
Prerequisite knowledge:
– Understanding of core Data Science methods, frameworks and libraries.
– An image of what Docker and Kubernetes are

Presentation Curriculum

Kubeflow, MLFlow and beyond - augmenting ML delivery
67:15
Hide Content