1:1 mentoring with Big Tech AI engineers
System Design18 Questions

Agentic System Design Interview Questions

Each question follows a structured format: problem restatement, clarifying questions, step-by-step architecture, Mermaid diagrams, and key tradeoffs. Practice the way you'd present in a real interview.

18Questions
25Companies
62Topics
2Free

Filters

Difficulty
Topics
Companies

18 questions

S10

Design a Proactive Personal Assistant Agent

Hard

Design a proactive personal assistant that manages a user's calendar and email — scheduling meetings, drafting replies, and surfacing important items before the user asks.

AgentsTool UseSafetyPrompt InjectionGoogle
S04

Design a Multi-Agent Orchestration System

Medium

Design a system where multiple specialized AI agents work together to complete a complex task — one researches, another drafts, a third critiques, and a fourth finalizes.

Multi-agentOrchestrationContext IsolationCoordinationTerminationMicrosoft
S19

Design a Conversational Commerce / Shopping Agent

Hard

Design a conversational shopping agent that understands customer needs through natural dialogue, recommends products, compares options, and helps complete purchases.

AgentsRAGE-CommerceSafetyAmazon
S14

Design a Customer-Support Triage & Resolution Agent

Hard

Design an AI agent that handles incoming customer support tickets end-to-end — triaging by urgency, resolving simple issues automatically, and escalating complex ones to humans.

AgentsTool UseRAGSafetyAmazon
S13

Design a Text-to-SQL / Conversational Analytics Agent

Medium

Design a system that lets non-technical business users ask questions in plain English and get accurate SQL queries and visualized results from a data warehouse.

RAGTool UseDataSecurityDatabricks
S02

Design a Deep Research Agent

Hard

Design an AI agent that takes a complex research question, autonomously searches the web, reads and cross-checks dozens of sources, and produces a well-structured report with proper citations.

ReActPlanningWeb ToolsCitationsLong-contextReflectionOpenAI
S15

Design an LLM Gateway / Model Router & Inference Proxy

Medium

Design an internal LLM gateway that every product team calls instead of hitting OpenAI, Anthropic, or Google directly.

InfrastructureCost OptimizationReliabilityRoutingAnthropic
S16

Design an Intelligent Document Processing & Workflow Agent

Medium

Design an intelligent document processing system that takes incoming documents (invoices, contracts, receipts), extracts structured data, validates it, and routes it downstream.

AgentsDocument AIAutomationComplianceStripe
S07

Design a Long-Term Memory System

Medium

Design a memory system that lets an AI assistant remember user preferences, facts, and past conversations across sessions and use them when relevant.

MemoryRetrievalSummarizationPersonalizationPrivacyForgettingOpenAI
S06

Design an Enterprise Knowledge Agent

Hard

Design an internal company assistant that answers employee questions by searching across Slack, Docs, Jira, and other tools — while respecting each user's access permissions.

RAGPermissionsConnectorsTool UseCitationsAuditMicrosoft
S17

Design a Trust & Safety / Content Moderation Agent at Scale

Hard

Design a trust and safety system that uses LLMs to moderate user-generated content on a platform with millions of posts per day.

ClassificationSafetyScaleCost OptimizationMeta
S12

Design an AI Coding Copilot — Real-Time Inline Completion

Hard

Design the inline code completion feature in modern IDEs — the gray-text suggestions that appear as you type — serving millions of developers in real time.

InferenceLatencyCost OptimizationDeveloper ToolsGitHub
S03

Design an Autonomous Coding (SWE) Agent

Hard

Design an AI agent that takes a GitHub issue, navigates a large codebase, writes a fix, runs tests, and opens a pull request — all without human help.

Code AgentsSandboxingTest-DrivenRepo NavigationVerificationGitHub
S09

Design an Agent Evaluation & Guardrail Platform

Hard

Design a platform that continuously evaluates AI agents in production and prevents them from taking unsafe actions.

EvaluationGuardrailsObservabilityRed-teamingCI/CDTracingAnthropic
S05

Design a Computer-Use / Browser Agent

Hard

Design an AI agent that controls a real web browser or desktop GUI to complete tasks on the user's behalf — filling forms, clicking buttons, and navigating websites.

Computer UseGUI GroundingVisionSafetyRecoveryInjection DefenseAnthropic
S11

Design an AI On-Call / Incident-Response Agent

Hard

Design an AI agent that helps on-call engineers respond to production incidents — investigating alerts, correlating logs and metrics, identifying root cause, and suggesting fixes.

AgentsTool UseObservabilityReliabilityGoogle
S08

Design a Real-Time Voice Agent

Hard

Design a voice AI agent that handles real phone calls with low latency — understanding speech, deciding what to say, and responding naturally with proper turn-taking.

StreamingLatencyTurn-takingBarge-inASR/TTSTool UseGoogle
S18

Design a Synthetic Data Generation & Curation Agent

Hard

Design a pipeline that generates high-quality synthetic training data at scale — for fine-tuning models, building evaluation sets, or augmenting sparse datasets.

DataEvaluationTrainingQualityAnthropic

Want to practice knowledge-based interview questions too?

30 Knowledge Questions