Platforms for Humans and Machines: Engineering for the Age of Agents
summary
Platform claims become testable specifics: Banking Circle runs Atlas with Kubernetes compute, exposes schema-validated APIs and optional MCPs, and requires local-first validation so agents iterate without breaking pipelines. Observability is API-accessible (logs metrics traces) so an agent can close the loop. Measure impact with DORA delivery metrics, support-request counts and reliability metrics. Contribution flow accepts agent-generated patches but enforces guardrails, CLAUDE.md and repository agents.md for maintainability.
AI-written summary · about
chapters (13)
- 0:00Platform engineering at Banking Circle
- 2:31Developer deployment struggles
- 5:42AI agents amplify pain points
- 7:09Self-service for AI-ready platforms
- 8:34API-first for humans and machines
- 10:19Local-first workflows for agents
- 11:22Observability for agent instructions
- 12:47Documentation for AI agents
- 14:12agents.md for agent docs
- 15:38Contributing to internal platforms
- 17:23Measuring AI platform success
- 19:29AI-ready platform principles
- 20:33Leveraging AI for best practices
related talks
38mPlatform Engineering in the Age of AI Teammates
33mHumans on the loop, not in it: Taking agentic engineering end to end
34mUnlock the full potential of Generative AI with Agent Platforms
31mClosing panel: The future of platform engineering and what to build next
01h 02mPlatform Engineering in the age of Generative AI- 33mHumans on the loop, not in it: Taking agentic engineering end to end
28mClosing panel: The future of platform engineering and what to build next
27mThinking in platforms: Thrive in the agentic era