Backend Engineer - Personalization

Spotify Spotify · Consumer · London, United Kingdom +1 · Personalization

Backend Engineer role focused on building and owning real-time recommendation and content discovery systems for Spotify's Now Playing View. The role involves integrating with ML serving infrastructure, ensuring low latency, and designing systems for agent-generated code, with a strong emphasis on A/B testing and production reliability.

What you'd actually do

  1. Own the backend systems that rank and serve personalized content to hundreds of millions of Spotify users on the Now Playing View, one of Spotify's most visible surfaces.
  2. Work in an AI-native development workflow. To implement backend changes across services, review their output, and ensure architectural coherence across the system.
  3. Integrate with ML serving infrastructure and collaborate with ML engineers to bring models from training to production within tight latency constraints.
  4. Design the patterns, tests, and guardrails that keep agent-generated code consistent, safe, and production-ready across services.
  5. Run and analyze A/B experiments from instrumenting new features to interpreting rollout metrics. Shipping what the data says works.

Skills

Required

  • backend systems ownership
  • ranking systems
  • serving personalized content
  • ML serving infrastructure integration
  • low latency constraints
  • A/B experiment analysis
  • production reliability
  • monitoring
  • alerting
  • incident response

Nice to have

  • machine learning
  • data science
  • research science
  • product management

What the JD emphasized

  • AI-native development workflow
  • agent-generated code

Other signals

  • backend systems that rank and serve personalized content
  • AI-native development workflow
  • integrate with ML serving infrastructure
  • bring models from training to production within tight latency constraints
  • design the patterns, tests, and guardrails that keep agent-generated code consistent, safe, and production-ready