Dev Interrupted Your engineers need an AI control plane, not more tools | Guild.ai’s James Everingham
45 snips
Mar 10, 2026 James Everingham, former Head of Dev Infra at Meta and CEO of Guild.ai, builds enterprise control planes for safe, auditable AI in development. He talks about weaving AI into the software lifecycle, developer-driven emergent agents like onboarding and risk scoring, why top-down mandates fail, and the need for a central AI control plane to govern, log, and scale agent workflows.
AI Snips
Chapters
Transcript
Episode notes
Move AI From Editor Into Source Control
- AI delivers more leverage when moved from the editor into infrastructure like source control.
- James explains source control is the canonical truth where intercepting tool connections lets agents do higher-impact tasks across teams.
Create A Centralized Agent Catalog
- Start with a centralized surface where engineers can discover, fork, and improve agents rather than building point solutions.
- James credits DevMate's catalog approach for making it easy for engineers to steal, adapt, and learn quickly.
Risk Score Agent Eliminated Year End Code Freeze
- An internal risk score agent helped Meta remove its year-end code freeze by estimating diff risk before landing.
- James describes the challenge framing “eliminate code freeze” led engineers to build the diff risk score solution.