Intermediate interview questions in Cloud Platforms, page 3
Idempotency in data ingestion pipelines
Idempotency means re-running a step yields the same result with no duplicates; it matters because retries and at-least-once delivery are inevitable; achieve it with deduplication keys or upserts.
Deploying a real-time inference endpoint
Package the model artifact and inference code in a container, choose instance type and autoscaling, configure the endpoint with health checks, and plan safe rollout like canary plus monitoring.
Inference performance bottlenecks on Lambda
Cold starts loading the model, memory and CPU limits, no GPU, and package size dominate; mitigate with provisioned concurrency, loading the model once outside the handler, smaller models, and right-sized…
How would you build CI/CD for an ML model?
Data and model versioning, automated training plus evaluation gates, model registry, deployment with monitoring and retraining triggers.
How would you speed up slow single-GPU training?
Vertical scaling to bigger or multi-GPU instances, then data-parallel or model-parallel distributed training across nodes.
Design a highly available web application
Redundant stateless instances across multiple zones behind a load balancer with health checks, auto-scaling, and a replicated multi-AZ datastore.
Explain the Strangler Fig pattern
A facade routes traffic, new services gradually replace legacy features one slice at a time, and the old system is retired when fully strangled.
Managed services versus self-hosting trade-offs
Managed services cut operational burden and speed delivery but cost more and limit control; self-hosting offers full control and tuning at the price of patching, scaling, and reliability work.
Balance agility and compliance in regulated cloud
PaaS for speed where allowed, IaaS where control is required, enforced by encryption, IAM least privilege, network isolation, policy-as-code guardrails, and continuous audit logging.
When to choose bare metal over a VM
Bare metal suits latency-sensitive or high-throughput workloads needing no hypervisor overhead, single-tenant isolation for compliance, or direct hardware and licensing access.
Configure a Kubernetes Horizontal Pod Autoscaler
HPA adjusts replica count toward a target CPU metric, needs the metrics server and pod resource requests, and scales a deployment between min and max.
Strangler Fig with serverless and an event bus
API Gateway acts as the routing facade, new features run as Lambda functions, an event bus decouples and fans out to new services, and traffic shifts feature by feature until the monolith…
Optimize cost of a big-data analytics platform
Storage tiering and lifecycle plus compression and partitioning; compute via spot instances, right-sizing, and efficient file formats; query and pipeline optimization to scan less data.
Event bus versus message queue for triggers
A queue is point-to-point buffered work for one consumer group; an event bus routes and filters one event to many decoupled subscribers. Event bus wins when many independent services must react.
Design a multi-tenant model serving platform
Share infrastructure to cut cost while enforcing tenant data isolation, fair resource allocation against noisy neighbors, and per-tenant performance via quotas and autoscaling.
Difference between metrics and logs
Metrics are aggregated numeric time series good for trends and alerting; logs are discrete timestamped event records good for detailed root-cause analysis.
High availability versus fault tolerance
HA minimizes downtime via redundancy and failover; fault tolerance survives failure with zero interruption.
Blue/green versus canary deployments
Blue/green flips all traffic between two full environments; canary shifts a small slice gradually while watching metrics.
Diagnosing slow auto-scaled PaaS workloads
Application metrics like request latency, throughput, and DB query time; infrastructure metrics like CPU, memory, and scaling lag.
Migrating a stateful monolith to the cloud
Assess and inventory, pick a migration pattern like rehost or replatform, handle data migration and cutover, mitigate downtime and data-loss risk.
We are hiring for this. Every open role lists the topics its interview covers, so you can prepare for the real thing rather than guessing.
See open roles