ML Systems & MLOps

Training schedules, drift, rollouts, experiments and the arithmetic of serving models.

Advanced · 14 problems · 0 passed · Background reading: MLOps — Roadmap

1. 01Deterministic Hash-Based Train/Test Splitmedium
2. 02Linear Learning-Rate Warmup Scheduleeasy
3. 03Cosine Annealing Learning Ratemedium
4. 04Early Stopping Patience Checkmedium
5. 05Training Batch GPU Memory Estimatormedium
6. 06Population Stability Index (PSI) for Drift Detectionmedium
7. 07Feature Skew Detection Between Training and Servingmedium
8. 08Canary Rollout Traffic Assignmentmedium
9. 09INT8 Weight Quantization From Scratchmedium
10. 10Parallel Sum Reduction (WebGPU)medium
11. 11Consistent Hash Ring Lookupeasy
12. 12A/B Test Statistical Significance (Two-Proportion Z-Test)hard
13. 13Page-Hinkley Concept Drift Statistic Updatehard
14. 14Pipeline DAG Topological Sorthard

Next track: Coding Interview Patterns