Design a geo-distributed A/B Testing System
Last updated: April 3, 2026
Quick Overview
Design a geo-distributed a/b testing system that handles millions of requests. Discuss trade-offs in consistency, availability, and performance.
NVIDIA
April 3, 202626
3
1,539 solved
Design a geo-distributed a/b testing system that handles millions of requests. Discuss trade-offs in consistency, availability, and performance.
ML system design at NVIDIA goes beyond model selection. This Onsite question evaluates your ability to design end-to-end ML pipelines, from data collection to model serving, while considering production constraints like latency and reliability.
What the Interviewer Expects
- Map the business problem to a concrete ML objective
- Propose reasonable features and a baseline model
- Discuss basic model evaluation metrics
- Outline a simple serving architecture
Key Topics to Cover
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
- How would you handle a 10x increase in prediction requests?
- How would you ensure fairness and reduce bias in the model?
- What would you do if model performance degrades over time?
- What is your model retraining strategy?
Practice a Similar Problem on Codemia
Solve a related problem with our interactive workspace, get AI feedback, and view detailed solutions.
Solve on CodemiaSample Answer
Requirements Clarification
Before diving into the architecture, clarify the scope with the interviewer. For geo-distributed A/B Testing System, key functional requirements inclu...
Capacity Estimation
Estimate the scale to drive design decisions. Assume 100M DAU with an average of 10 actions per user per day = 1B requests/day ~ 12K QPS average, ~36K...
Submit Your Answer
