Difficulty Distribution
Beginner0
Intermediate0
Advanced2
Skills You'll Practice (6)
All 2 Scenarios
Click any scenario to practice it. No account required.
advanceddesign
Deploy LLM to Production
Your team fine-tuned a Llama-3 70B model for customer support. It needs to serve 10K requests/minute with <2 second latency. Walk me through: infrastructure choices, model serving, cost optimization, monitoring, and fallback strategy.
LLM deploymentMLOpsGPU infrastructure
30 min
advancedincident
Training Pipeline Breaking Down
Your ML training pipeline (Python + PyTorch + Kubernetes) has been running for 8 hours and just crashed. It's supposed to train a computer vision model on 50TB of data. The crash was OOM on a specific GPU node. You have a demo tomorrow. Walk me through recovery.
Training infrastructureGPU debuggingRecovery
20 min