Software Engineer, Communication Safety

Roblox Roblox · Consumer · San Mateo, CA · Software Engineering

Software Engineer on the Communication Safety team at Roblox, focusing on detecting and mitigating unsafe communication at massive scale. The role involves working at the intersection of backend infrastructure and machine learning, ensuring safety systems are accurate, performant, and reliable. Key responsibilities include designing and coding features for the core safety platform, improving pipeline performance, and collaborating with product and partner teams. Experience with ML model lifecycle, specifically model serving, experimentation, and performance monitoring, is required.

What you'd actually do

  1. Design, code, and test innovative features for our core safety platform, navigating the full development lifecycle from initial design to production deployment.
  2. Improve the performance, memory usage, and stability of our pipeline across all content types including chat game assets and in game experience.
  3. Work closely with product, design, and partner teams to maintain high execution quality and pace.
  4. Be mentored by senior engineers and actively engaged in architectural discussions and system design for our text safety solutions.

Skills

Required

  • C#
  • Python
  • building and maintaining scalable, high-availability backend services
  • lifecycle of machine learning models
  • model serving
  • experimentation
  • performance monitoring
  • handling production incidents
  • improving observability and recovery times in distributed systems
  • collaborating cross-functionally with Product stakeholders

Nice to have

  • 1-3 years of experience in Software Engineering/Computer Science

What the JD emphasized

  • real-time moderation
  • massive scale
  • low-latency execution
  • machine learning models
  • model serving
  • performance monitoring
  • production incidents
  • observability
  • recovery times

Other signals

  • real-time moderation
  • massive scale
  • low-latency execution
  • machine learning models
  • model serving
  • performance monitoring