Missions
Each mission drops you into a simulated cluster with a real-world problem, a terminal and live dashboards. Objectives check themselves. Scores reward speed and fewer hints, and every run reports its energy, cost and carbon.
Two Nodes, One Job
Your first multi-node training job won't even submit. Fix the batch script.
Fallen Off the Bus
A training job died overnight. Find the bad GPU, quarantine the node and get the run going again.
The Slow Lane
Throughput just halved across 32 GPUs, but nothing failed. Find the straggler.
Feed the Beast
16 H100s running at 26% MFU. Tune the training config past 55% without running out of memory.
Watts Up
The facility asks for a 20% power cut. Keep at least 85% of the throughput.
Night Shift
A GPU is busy with no job scheduled. Investigate, contain and remove persistence.
Seventy Billion Reasons
A 70B model on 32 GPUs: it doesn't fit. Choose a parallelism strategy that does, and runs fast.
Build the Pod
Spec a GPU pod that trains a 13B model in two weeks, within budget and the facility's power envelope.