About the role
You keep personal agents useful after the first demonstration, which means making a fleet of household computers maintainable without treating people's private data as telemetry. That constraint is the interesting part: you have to operate what you cannot look inside.
The work
Build health checks, progressive updates, rollback, resource monitoring and incident runbooks. Separate device health from content. Design support access that users can authorize for a limited purpose and revoke, and test recovery from power and network failures.
What good looks like
In your first 90 days, establish a pilot fleet dashboard, staged rollout process and three reproducible failure drills.
Evidence we look for
Bring experience operating distributed services or connected devices. Show how you choose service objectives, contain incidents and reduce repeated support burdens.
What we need to see
- You have operated distributed services or connected devices in production
- You choose service objectives deliberately and can explain why those and not others
- You contain incidents and then remove the class of problem rather than the instance
- You can debug a fleet with aggregate signals only, without reaching into private data
Nice to have
- Consumer hardware fleets specifically
- Privacy-preserving telemetry or differential privacy
- You have reduced a support burden measurably
The exercise
Plan an update rollout where a small number of devices boot successfully but fail to reopen their encrypted vaults.
Where and how we work
In the office together five days a week, in any of these cities. Remote-friendly around your family, arranged one person at a time.