Low(er) Code ML Pipelines with Conduit
summary
Conduit is a Python package plus a terminal UI that claims to cut time-to-production by templating repos, letting teams paste notebook code into templates with no refactor, assembling pipelines from a component library and generating Kubeflow pipelines for Vertex AI. It scaffolds server files, creates versioned models in Vertex, deploys scalable prediction endpoints (demo reports ~300 ms latency), and exposes editable code so pipelines and endpoints are runnable and customizable rather than just theoretical.
AI-written summary · about
chapters (5)
- 0:00Low-code ML pipelines intro
- 3:12ML model production challenges
- 8:50Introducing Conduit and terminal UI
- 11:03Why Conduit matters for data scientists
- 13:30Q&A
related talks
26mBuilding Portable ML Pipelines for Development and Production
36mDemocratization of ML Pipelines and Bringing ML Workflows to Heterogeneous...
32mMLOps at Renault: A Generic Pipeline for Scalable Deployment
27mKubeflow for enabling AI Powered Drug Discovery and development...
05mCI/CD Pipelines and Machine Learning
45mVertex AI: Pipelines for your MLOps workflows
29mML Engineering Kubeflow
35mOrganizing data engineering around Kubeflow