skip to content

FAQ

For candidates

Is AI cheating here?
No. AI is allowed and graded. The transcript shows whether you drove it or it drove you.
What if I break it worse?
That's the blast-radius score, and recovering from your own mistakes counts for you, not against you.
What's recorded?
Terminal commands and files, never your screen, camera, or mic. See the privacy policy for the full list.
Can I practice first?
Free sessions are exactly that.
Do I bring my own AI, or do you provide one?
Today, bring your own, any tool, any model, openly allowed. A provided in-session assistant with graded prompt trail is on the v2 roadmap; your own AI stays allowed either way.

For employers

How do I pick which labs to send? I am not a DevOps engineer.
You have two paths. Browse the lab pool and filter by tool or difficulty, or upload the job description and the AI engine suggests the labs that actually test it. You can add preferences (must-cover tools, number of labs, time box per lab) and the suggestions respect them. Review, adjust, send.
How is this graded if an AI is involved?
The AI evaluator monitors the entire session, every command, exit code, and approach, and generates the graded report: diagnosis quality, recovery, blast radius. Pass/fail itself is anchored by deterministic scripts that check the system actually recovered, so a session still grades even if the AI evaluator is unavailable.
What about scenario leakage?
Randomized variants: one scenario, many faults. A leaked walkthrough is worthless the moment it is posted.
Do you proctor?
No, replay beats surveillance. You can watch the entire session, command by command.
What stacks are supported?
The whole DevOps surface: Kubernetes, Docker, Linux, Terraform, and monitoring (Prometheus/Grafana) at launch. Databases, networking, CI/CD, and cloud-provider scenarios are on the roadmap.
Why is there no screen or video recording?
Deliberate. The terminal replay already shows every command, exit code, and dead end, searchable text at kilobytes instead of video at gigabytes, with nothing from the candidate’s camera, mic, or screen ever captured. Video adds surveillance, not signal.
Can we use this to train the team we already have?
Team incident drills, scheduled game days on the same scenarios, with per-engineer reports, are on the v2 roadmap. Email us if you want in early.