Leveraging Agent Skills for Ray Workloads
summary
Anyscale ships an agent-native CLI plus SDK and MCP hooks that let coding agents (cloud code, cursor, CodeX, GitHub Copilot) run locally to spin up Ray workspaces, upload validated templates, deploy models (LLaMA example) and fine-tune Nvidia Cosmos 3 via a fully autonomous loop that inspects logs, suggests trade-offs, applies guarded fixes, and resubmits jobs. Skills use templates validated on Anyscale to cut hallucination, include pre-run hooks to block destructive commands, and report success rates as the QA metric.
AI-written summary · about
chapters (5)
- 0:00Introduction to talk
- 1:09Why build agent skills
- 8:43Deep dive into skills
- 20:49Lessons learned building skills
- 30:25Q&A and closing
related talks
29mBuilding Scalable AI Agents with Ray
31mManaging Large-Scale Ray Deployments: Cloud, On-Prem, and Kubernetes
21mHow Ray and Anyscale Make it Easy to do Massive-scale ML on Aerial Imagery
20mThe Quick Journey to Using Ray: How We Implement Ray and Anyscale to Speed up our ML Processes
13mRay and Anyscale: An Optimization Journey
42mLLMs on Autopilot: Running AI Agents on Kubernetes With Open Source Tools
40mAgents in Action: Advancing Open Source AI Missions
27m[Opening Keynote] Anyscale Demo: Machine Learning Application from Dev to Prod