Dataford
Popular roles
Software EngineerData AnalystData ScientistData EngineerBusiness AnalystAI EngineerMachine Learning EngineerProduct Manager
Browse
Browse All RolesEvery role hub, from analyst to MLBrowse All CompaniesCompany-specific interview loopsAll Interview GuidesThe full guide library
Top questions by role
Software EngineerData AnalystData ScientistData EngineerBusiness AnalystAI EngineerMachine Learning EngineerProduct Manager
Top questions by skill
SQLPythonStatisticsMachine LearningA/B TestingSystem DesignGenerative AIProduct SenseMetricsBehavioral
Browse all questions →Try a mock interview
ExperiencesCoaching
New
Practice
Mock InterviewsTimed interview simulations with feedbackSuccess PathYour 6-week structured planModulesCurated lessons by topicWebinarsTalks from ex-Big Tech data leadsPlaygroundA free-form scratch editor
Learn
BlogInterview strategy and career adviceTech Job Market ReportHiring trends across data and AI rolesFor UniversitiesDataford for career centersAbout DatafordWho we are and how we build
Pricing
Get started

Design WhatsApp's ML Messaging System

HardSystem Design00:00
Practice interviewer
In session
5 left
00:00

Whenever you're ready, walk me through how you'd approach this.

You need to log in / sign up to chat or submit.

Problem

Design WhatsApp.

Asked in the system design 1 stage. Treat this as an open-ended ML system design problem: first clarify which WhatsApp capability is in scope, then propose an end-to-end design for a high-scale messaging product.

Address:

  1. Functional and non-functional requirements and scale assumptions.
  2. Candidate generation, ranking, or classification stages as appropriate.
  3. Online versus batch inference, feature storage, and latency budgets.
  4. Training data, delayed labels, offline and online evaluation.
  5. Monitoring, privacy, training-serving skew, drift, and failure modes.

Assume end-to-end message interaction latency must remain low, user data must be protected, and the design must degrade safely when ML services are unavailable.