PracHub
QuestionsLearningGuidesInterview Prep
|Home/System Design/Assorted Health

Design a Conversational Flight Search, Booking, and Price-Tracking Service

Last updated: Jul 22, 2026

Quick Overview

Design a conversational flight service that turns vague travel intent into grounded multi-leg search, safe booking, and useful price alerts. Cover clarification, authoritative supplier data, repricing, idempotent payment flows, shared price watches, quotas, privacy, and graceful degradation at large scale.

  • medium
  • Assorted Health
  • System Design
  • Software Engineer

Design a Conversational Flight Search, Booking, and Price-Tracking Service

Company: Assorted Health

Role: Software Engineer

Category: System Design

Difficulty: medium

Interview Round: Onsite

# Design a Conversational Flight Search, Booking, and Price-Tracking Service Design a service for roughly 30 million daily active users. A user can chat with an agent to search for and schedule a flight, and can follow the lowest price for an itinerary. Queries may be underspecified, such as "I want to visit Europe next year," and valid results can include connecting flights rather than only direct routes. Explain how the conversational layer turns vague intent into a validated search, how itinerary search and booking remain correct, and how price tracking notifies users without overwhelming flight suppliers. ### Constraints & Assumptions - Supplier prices and availability can change between search and booking. - The language model is not a source of truth for schedules, fares, or booking state. - A booking or payment request may be retried. - Exact peak load, supplier quotas, and freshness targets must be elicited rather than assumed. ### Clarifying Questions to Ask - Does "schedule" include purchasing, or only selecting and saving an itinerary? - Which origins, date flexibility, budget, trip length, and destination preferences are known? - What does "lowest price" mean across cabin, baggage, refundability, and connection count? - How quickly must price alerts react, and how often may a user be notified? ### What a Strong Answer Covers - Structured intent extraction plus clarification for missing constraints - A bounded tool interface that grounds the model in authoritative search results - Multi-leg itinerary generation, ranking, repricing, and booking idempotency - Scalable caches, asynchronous price checks, deduplication, and supplier rate limits - Safety, privacy, observability, and graceful degradation ### Follow-up Questions - How do you stop the agent from inventing an airport or fare rule? - How would you search a continent without querying every airport equally? - What happens when the price rises at checkout? - How do you prevent millions of identical price watches from triggering identical supplier calls?

Quick Answer: Design a conversational flight service that turns vague travel intent into grounded multi-leg search, safe booking, and useful price alerts. Cover clarification, authoritative supplier data, repricing, idempotent payment flows, shared price watches, quotas, privacy, and graceful degradation at large scale.

|Home/System Design/Assorted Health

Design a Conversational Flight Search, Booking, and Price-Tracking Service

Assorted Health logo
Assorted Health
Jan 14, 2026, 12:00 AM
mediumSoftware EngineerOnsiteSystem Design
1
0

Design a Conversational Flight Search, Booking, and Price-Tracking Service

Design a service for roughly 30 million daily active users. A user can chat with an agent to search for and schedule a flight, and can follow the lowest price for an itinerary. Queries may be underspecified, such as "I want to visit Europe next year," and valid results can include connecting flights rather than only direct routes.

Explain how the conversational layer turns vague intent into a validated search, how itinerary search and booking remain correct, and how price tracking notifies users without overwhelming flight suppliers.

Constraints & Assumptions

  • Supplier prices and availability can change between search and booking.
  • The language model is not a source of truth for schedules, fares, or booking state.
  • A booking or payment request may be retried.
  • Exact peak load, supplier quotas, and freshness targets must be elicited rather than assumed.

Clarifying Questions to Ask Guidance

  • Does "schedule" include purchasing, or only selecting and saving an itinerary?
  • Which origins, date flexibility, budget, trip length, and destination preferences are known?
  • What does "lowest price" mean across cabin, baggage, refundability, and connection count?
  • How quickly must price alerts react, and how often may a user be notified?

What a Strong Answer Covers Guidance

  • Structured intent extraction plus clarification for missing constraints
  • A bounded tool interface that grounds the model in authoritative search results
  • Multi-leg itinerary generation, ranking, repricing, and booking idempotency
  • Scalable caches, asynchronous price checks, deduplication, and supplier rate limits
  • Safety, privacy, observability, and graceful degradation

Follow-up Questions Guidance

  • How do you stop the agent from inventing an airport or fare rule?
  • How would you search a continent without querying every airport equally?
  • What happens when the price rises at checkout?
  • How do you prevent millions of identical price watches from triggering identical supplier calls?

Submit Your Answer to Earn 20XP

Sign in to leave a comment

Loading comments...

Browse More Questions

More System Design•More Assorted Health•More Software Engineer•Assorted Health Software Engineer•Assorted Health System Design•Software Engineer System Design

Your design canvas — auto-saved

PracHub

Master your tech interviews with 9,000+ real questions from top companies.

Product

  • Questions
  • Learning Tracks
  • Interview Guides
  • Resources
  • Premium
  • For Universities

Browse

  • By Company
  • By Role
  • By Category
  • Topic Hubs
  • SQL Questions
  • AI Coding Questions
  • Compare Platforms
  • Discord Community

Support

  • support@prachub.com
  • (916) 541-4762

Legal

  • Privacy Policy
  • Terms of Service
  • About Us

© 2026 PracHub. All rights reserved.