Recommended preparation
Premium
Real-time Model Training Dashboard
Try the prompt first
Senior20 min first pass
Design a model-training progress dashboard with resumable metrics, run comparison, chart downsampling, stale-state signaling, failure diagnosis, and accessibility.
Constraints
- Keep client, server, and rendering ownership explicit.
- Cover loading, failure, recovery, and accessible interaction states.
Challenge summary
Design a real-time training dashboard that merges run lifecycle, metrics, and logs without losing terminal evidence during bursts or reconnects.
What you'll practice
- Bridge a coherent snapshot to a sequenced, resumable event stream.
- Merge lifecycle, metric, and log fields without allowing late samples to reopen a run.
- Bound chart and log rendering while preserving gaps and terminal events.
- Provide comparable units, freshness, diagnosis context, and accessible chart alternatives.
What Premium unlocks
Premium unlocks the late-metric example, contiguous resume model, chart trade-offs, and terminal-state diagnosis.
Free related practice
Upgrade to FrontendAtlas Premium to access this system design scenario. Already upgraded? Sign in to continue.