Keynote: A Tale of Using Kubeflow to Make the Electricity Smarter in China
summary
Deploys Kubeflow on Kubernetes to unify TensorFlow and PyTorch pipelines, PV/PVC storage, and orchestrated distributed training; connects to State Grid of China data to run OCR and time‑series models for invoices, load forecasting, equipment vision and hydropower monitoring. Handles a client case with 400 GPUs and tens of concurrent models, adds CI for fast iteration, automated resource reallocation, experiment tracking and a multidimensional dashboard. Claims faster model iteration and production deployment and shows a working AI OS used in finance and operations.
AI-written summary · about
chapters (6)
- 0:00Smarter electricity in China
- 4:26AI adoption challenges in power
- 9:25Kubeflow as solution
- 13:30Kubeflow features for AI
- 16:49AI OS in power generation
- 19:41Demo and system integration
related talks
15mKeynote: Machine Learning on Kubernetes Made Easy With Kubeflow
22mKeynote: Kubeflow ML on Kubernetes
10mKeynote: Machine Learning using Kubeflow and Kubernetes
21mIntroducing Kubeflow (w. Special Guests Tensorflow and Apache Spark)
36mIntroducing Kubeflow (w. Special Guests Tensorflow and Apache Spark)
46mDeploying scalable Machine Learning models with Kubernetes
54mEconomics and Best Practices of Running AI/ML Workloads on Kubernetes
36mIntroducing Kubeflow (w. Special Guests Tensorflow and Apache Spark)