AI · SRE· 2026 · Creator
ApeironOps — Autonomous Incident Commander
Autonomous incident commander for verified software recovery: Microsoft Foundry-assisted agents that investigate incidents, block unsafe fixes, validate recovery, and generate postmortems.
- Agents investigate incidents and propose fixes
- Unsafe fixes are blocked before they ship
- Recovery is validated, then postmortems are generated
Context & Goals
ApeironOps applies the "verified recovery" idea to incident response: instead of an AI agent that pushes whatever fix it thinks of, agents investigate the incident, propose remediations, and pass them through safety checks before anything reaches production. The system validates that recovery actually happened and writes the postmortem afterward.
How it works
- Investigation agents — Microsoft Foundry-assisted agents gather signals and reconstruct what went wrong.
- Safety gating — proposed fixes are checked and unsafe ones are blocked.
- Verified recovery — the system confirms the incident is actually resolved before closing it.
- Postmortems — a write-up is generated from the investigation trail.
Reflection
This project extends the philosophy behind ApeironCode — autonomy with control — from developer workflows into SRE and reliability engineering.
Up next
Echo — Real-Time AI Speaking Coach
Real-time AI speaking coach with live transcription, filler-word detection, eye-contact tracking, mock interviews, and Gemini-powered feedback reports.