Building Cloud Native Agentic Workflows on Kubernetes for Preventative Healthcare

Explore how to build cloud-native agentic workflows on Kubernetes for preventative healthcare applications in this 22-minute conference talk from KubeCon + CloudNativeCon. Learn to automate entire healthcare outreach loops using an open-source stack including Kubernetes, APISIX, and Prometheus combined with the AutoGen multi-agent framework. Discover how to define U.S. Preventive Services Task Force screening criteria, filter patient records, and generate personalized emails using open-source LLMs like Llama 3 and DeepSeek-R1 served through OpenAI-compatible APIs on Kubernetes-native AI accelerators. Master the YAML and Helm chart workflows that maintain model endpoints, agents, and Streamlit front-end deployments. Gain practical skills in building model inference endpoint orchestration layers with Kubernetes and Helm charts, managing API traffic and authentication using APISIX built on NGINX and etcd, connecting agents through asynchronous Python programming, and implementing comprehensive monitoring and observability solutions with Prometheus and Grafana for scalable preventative healthcare systems.