ML Systems & Infrastructure
ML Systems & Infrastructure 1
Fifteen Python exercises that make you reason about the systems artefacts an ML platform runs on: Docker layer caches and image sizes, Kubernetes scheduling and rollouts, CIDR planning and firewall rules, SLO error budgets and tail latency, CI critical paths, retry backoff and spot-instance economics.
50
problems
5
topics
15
free
01
Container Images & Builds
10 problems- 1Multi-Stage Build Size ReductionFREEEasy
- 2Effective Environment After Layered ENV InstructionsPROEasy
- 3Total Image Size from Unique LayersPROEasy
- 4Resolve an Image Tag from a Semver ConstraintFREEMedium
- 5First Cache-Invalidated Build LayerPROMedium
- 6Highest Semver Matching a Caret RangePROMedium
- 7Layer Reuse Ratio Across an Image SetPROMedium
- 8Order Build Stages by DependencyPROMedium
- 9Dockerfile Layer Cache InvalidationFREEHard
- 10Minimum Rebuild Cost with Layer CachingPROHard
02
Kubernetes Scheduling & Rollouts
10 problems- 11Horizontal Pod Autoscaler Replica CountFREEEasy
- 12Desired Replicas from CPU UtilizationPROEasy
- 13Does a Pod Fit on a NodePROEasy
- 14Rolling Update Surge and Availability BoundsFREEMedium
- 15Rolling Update Pod Availability BoundsPROMedium
- 16Bin-Pack Pods onto Nodes (First-Fit Decreasing)PROMedium
- 17Filter Nodes by Taints and TolerationsPROMedium
- 18Spread Pods Across Zones for Anti-AffinityPROMedium
- 19Bin-Pack Pods onto NodesFREEHard
- 20Minimum Rollout Steps Under Surge and UnavailabilityPROHard
03
Networking & Addressing
10 problemsSign up free to see all 50 problems
Create a free account to see the full ML Systems & Infrastructure 1 list, track your progress, and save problems to lists.
Sign up free