ByteGeist Support Lab

ByteGeist Support Lab helps learners practice the reasoning behind help-desk and cloud-support work. Each incident presents a user problem, an investigation environment, and a place to explain a diagnosis and proposed fix.

The core rule is AI teaches. Code grades. The coach guides the investigation; deterministic code controls the score and scenario state.

1. Choose an incident from the queue.
2. Run supported troubleshooting commands in the simulated terminal and collect evidence.
3. Use the Troubleshooting Buddy for questions, progressive hints, or an explanation of a selected result.
4. Write a root-cause diagnosis and proposed resolution.
5. Submit the resolution for deterministic grading and review the feedback.

• The Missing File Server: distinguish DNS resolution failures from general network connectivity.
• Access Denied: investigate group membership and file-share access.
• AccessDenied in S3: separate authentication from IAM authorization.
• Service Won’t Start: investigate Linux service state and permissions.
• Server Is Up, App Is Down: distinguish infrastructure reachability from application availability.

The Troubleshooting Buddy offers one supported simulator check at a time, explains why it matters, and asks the learner to interpret the result. A command can be loaded into the terminal, but the learner executes it. Guided progress follows collected evidence from the session transcript.

The frontend uses React. The application uses AWS services and Amazon Bedrock for coaching. Grading remains separate from generated coaching responses.

• Designing troubleshooting workflows around evidence and reasoning.
• Connecting a frontend to server-side scenario and coaching services.
• Keeping deterministic scoring separate from generated explanations.
• Presenting command output, diagnosis, and feedback in a usable interface.

The terminal is a simulation. Completing an incident demonstrates work within that training scenario, not a repair to a real production system.