Free Scenario Question Bank

Designed by-Priya Dw | Authored by Experienced AI/ML Architects | Reviewed by relevant & experienced tech panels

Scenario Bank: Sample Questions with Answer Guidance

AI/ML - Drift Investigation: A model drift alert spikes after a product launch. Which first action best isolates root cause?

Answer approach: Segment telemetry by new vs baseline traffic first, then compare feature distribution and label lag by segment before changing model parameters.

Cloud Architect - Resilience vs Cost: Multi-region failover improves RTO but raises cost. Which architecture pattern balances both?

Answer approach: Use pilot-light or warm-standby where business critical paths stay hot and lower-priority workloads scale on-demand after failover.

Data Engineer - Late Events: Late-arriving events break daily aggregates. Which ingestion strategy preserves correctness?

Answer approach: Apply event-time watermarks and incremental backfill windows so historical partitions are corrected without full reload.

Security - Least Privilege at Scale: A new workload needs least privilege while preserving developer velocity. Which control plane design is most effective?

Answer approach: Implement role templates with short-lived credentials, automated policy linting, and mandatory approval for privilege escalation.

Additional Scenario Prompts

GenAI Governance: A team wants to use public LLM APIs for support tickets. How do you reduce data leakage risk while keeping response quality?

Platform SRE: Build times increase after adding security scans. Which CI design keeps compliance checks while reducing delivery delay?

Data Platform: A medallion architecture is in place but BI dashboards still disagree with finance reports. Which data contract controls should be introduced?

Identity Security: Contractors require temporary production access. Which access design offers strongest auditability and minimum standing privilege?

FinOps: GPU training spend doubled this month. Which first triage metrics best identify waste before model-level optimization?

Multi-Cloud: A latency-sensitive workload spans AWS and GCP. Which network and data-consistency strategy avoids split-brain risk?

Compliance: New privacy policy mandates regional data residency. Which architecture modifications should be prioritized first?

Incident Response: A suspicious token reuse pattern appears in logs. Which sequence should security follow in the first 30 minutes?

Practice Next

AI/ML Hub | Security Hub | Data Engineer Hub | Architect Hub

GCP | AWS | Azure | Databricks | ISC2 | Free Coupon Page