Dataford
Top 50
Topic roadmap
Updated weekly · Last refresh Sep 13

Top 50 computational cost Interview Questions

The most frequently asked computational cost questions across all roles and companies, ranked by real interview frequency. Updated daily.

50questions
~7htotal time
21companies covered
Track your progressSign up free to work through all 50 questions and resume where you left off.
Start practicing free →
1
Feature DriftStart here. 4 questions · ~32 min
Past Model Choice RationaleHard

Explain why a past ML model beat simpler alternatives, including trade-offs, evaluation, and production risks.

Feature Driftcomputational costfailure modesAmazon ServicesAmazon
More Feature Drift questions with a free account
2
gpu hardware9 questions · ~72 min
Transformer Memory vs ComputeHard

Explain why transformer inference is often memory-bound on modern GPUs and how to reduce bandwidth pressure.

gpu hardwaremodel inferencecomputational costAdobe
Identify Compute vs Memory BatchsizesHard

Design a practical benchmark to establish whether each batch size is compute-bound or memory-bound.

gpu hardwareml inferencecomputational costAmazon
More gpu hardware questions with a free account
3
ml inference5 questions · ~40 min
Handle Operations Under Limited MemoryMedium

Design algorithms and operations that process ML workloads reliably when available memory is limited.

ml inferencecomputational costoffline syncDatadog
More ml inference questions with a free account
4
latency6 questions · ~48 min
Self-Attention MechanismHard

Implement numerically stable self-attention with masking, batching, and clear complexity guarantees.

latencyml inferencecomputational costMeta
More latency questions with a free account
5
inference latency9 questions · ~72 min
More inference latency questions with a free account
6
distributed training5 questions · ~40 min
More distributed training questions with a free account
7
More topics12 questions · ~96 min
Design a Latency Aware Reasoning AgentMedium

Design an agentic assistant that decides when to use deeper reasoning versus fast responses, while managing latency, cost, and quality.

iterative reasoninginference latencycomputational costGeneral Dynamics Information TechnologyOpenAIAAccso
Vector Search for Massive DatasetsHard
Recently asked

Design a distributed vector search index that supports massive collections while maintaining low query latency and reliable freshness.

Vector Searchlow latencydistributed systemsAAgileengineGrafana LabsReply
LLMs vs Traditional NLPHard

Design an LLM-based NLP platform for enterprise search, summarization, and extraction at scale with low latency and strong monitoring.

factual groundingModel Servingcomputational costJPMorganChase
More questions with a free account

Sign up to see every question

Create a free account to unlock this list and practice real interview questions.

Get my prep plan
The finish line: interview-readyComplete all 50 questions to finish this plan.