Design a distributed Ad Serving System

Last updated: September 2, 2025

Quick Overview

Design a distributed ad serving system that handles millions of requests. Discuss trade-offs in consistency, availability, and performance.

Snowflake

System Design

Machine Learning Engineer

Snowflake

September 2, 2025

Machine Learning Engineer

Onsite

System Design

Hard

4,386 solved

Design a distributed ad serving system that handles millions of requests. Discuss trade-offs in consistency, availability, and performance.

This is a common system design question asked during Onsite at Snowflake. The interviewer expects you to demonstrate your ability to design large-scale distributed systems, make well-reasoned trade-offs, and communicate your thought process clearly. Snowflake values engineers who can think about scalability from day one.

What the Interviewer Expects

Drive the design discussion proactively with minimal interviewer guidance
Perform detailed capacity estimation and use it to inform design decisions
Design for global scale with multi-region deployment and data consistency
Deep dive into 2-3 critical components with implementation-level detail
Address complex trade-offs: CAP theorem, eventual consistency, conflict resolution
Discuss operational excellence: deployment strategy, chaos engineering, SLOs/SLIs
Propose a phased rollout plan from MVP to full-scale system

Key Topics to Cover

Monitoring, logging, and alerting

Requirements gathering and capacity estimation

High-level architecture and component design

Message queues and async processing

Load balancing and horizontal scaling

Partitioning and sharding strategies

How to Approach This

Start by clarifying functional and non-functional requirements with the interviewer.
Estimate the scale: QPS, storage, bandwidth. This drives your design decisions.
Draw a high-level architecture first, then deep dive into 1-2 critical components.
Discuss trade-offs explicitly (e.g., consistency vs availability, SQL vs NoSQL).
Address failure scenarios, monitoring, and how the system handles 10x traffic spikes.

Possible Follow-up Questions

What would the deployment pipeline look like for this system?
How would you handle a 10x increase in traffic overnight?
How do you ensure data consistency across multiple services?

Practice a Similar Problem on Codemia

Solve a related problem with our interactive workspace, get AI feedback, and view detailed solutions.

Solve on Codemia

Sample Answer

Requirements

Functional Requirements

Ad Serving: Serve ads to users based on real-time targeting criteria, including demographics, location, and user behavior.
Analytics: Provide real-time analyt...

Capacity Estimation

To estimate capacity, let's assume:

Daily Active Users (DAUs): 10 million users
Ad Requests per User: 5 requests per session
Sessions per User per Day: 2 sessions

Calculation

Submit Your Answer

Markdown supported

Snowflake Machine Learning Engineer Interview Guide

Interview process, tips, and preparation timeline