Meta·Software Engineer·Onsite - System Design / Architecture·Senior

SeniorPrefer not to say

Jun 2026

Summary

Meta system design round for a software engineer role. One big open-ended question about building a coding platform at scale, which sounds straightforward until you realize how many moving parts they actually want you to cover.

Questions Asked (1)

Design a coding platform like LeetCode that supports large-scale code execution and a live global leaderboard. Cover how you'd handle test case storage, orchestrate code execution, make the message queue fault tolerant, and scale the whole thing.

System DesignTechnical Trade-offsData Modeling

Author's notes

I started with the leaderboard because it felt like the flashiest part and I figured I'd impress them.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Start by clarifying requirements and scale, then propose a high-level architecture that separates submission handling, code execution, and leaderboard updates. Dive into each component (test case storage, execution orchestration, message queue fault tolerance, scaling) with specific technologies and trade-offs, emphasizing reliability and low latency.

Pro tip: Emphasize idempotency and exactly-once processing in the execution pipeline to avoid duplicate submissions and leaderboard corruption, and discuss how you'd monitor and alert on queue lag and execution failures.

1. Clarify Requirements and Scale

Ask about expected number of users, submissions per second, supported languages, and leaderboard update frequency. Define SLAs for execution latency and leaderboard consistency.

2. High-Level Architecture

Sketch components: API gateway, submission service, message queue, execution workers, test case storage, results database, leaderboard service, and caching layer. Explain data flow from submission to result.

3. Test Case Storage and Execution Orchestration

Propose storing test cases in object storage (e.g., S3) with metadata in a database, and orchestrating execution via a queue-based system where workers pull tasks, run code in sandboxed containers, and report results.

4. Message Queue Fault Tolerance and Scaling

Design the queue for durability, replication, and exactly-once processing using acknowledgments, dead-letter queues, and idempotent consumers. Scale horizontally by adding workers and partitioning the queue.

5. Leaderboard and Global Scaling

Implement a real-time leaderboard using a distributed cache (e.g., Redis sorted sets) with periodic persistence to a database. Scale globally with regional deployments, CDN for static assets, and eventual consistency for leaderboard updates.

Key Points to Mention

Use of containerization (Docker) and orchestration (Kubernetes) for isolated, scalable code execution.
Message queue selection (e.g., Kafka, RabbitMQ) with replication, partitioning, and dead-letter queues for fault tolerance.
Idempotent submission processing and exactly-once semantics to prevent duplicate executions.
Test case storage in object storage with caching for frequently used cases, and versioning for updates.
Leaderboard implementation with Redis sorted sets for O(log N) updates and reads, and periodic snapshots to a database.
Global scaling via multi-region deployment, read replicas, and CDN for static content, with eventual consistency for leaderboard.

AI-generated suggestions, not part of the candidate's original notes. May be inaccurate — verify before relying on them.