Coding Chats

Ona - the AI software engineer that works while you sleep.

Feb 26, 2026
Matt Boyle, Founder/CEO building Ona, an AI software-engineer platform. He explains autonomous AI agents that parallelize work across isolated cloud dev environments. They discuss planning mode that turns tickets into specs, enterprise security inside customer VPCs, managed model supply, and automations that scan, fix, and open pull requests while you sleep.
Ask episode
AI Snips
Chapters
Transcript
Episode notes
ADVICE

Plan Thoroughly Before Parallelising Agents

  • Use a planning mode that asks questions and creates a specification before parallelising work.
  • Ona pulls context from monorepos, Jira/Linear/Notion and maps dependencies so multiple cloud environments can run independent tasks safely.
ADVICE

Run Agents Inside Customer VPCs For Security

  • Run agent environments inside the customer's cloud VPC to reduce security risk and retain network control.
  • Ona deploys into customers' VPCs with guardrails, one‑click contractor offboarding, and zero data retention guarantees with model providers.
INSIGHT

Model Supply Is A Managed Service Problem

  • Treat model supply like a managed service rather than 'bring your own key' because model quality and reliability shape product experience.
  • Ona provides managed models with failover across providers to keep agent behavior consistent.
Get the Snipd Podcast app to discover more snips from this episode
Get the app