Rockstar 3

Rockstar 3

Backend Software Engineer (ML Infra)

San Francisco, California, United States

Sponsorship not specifiedDetected 213 days ago
PythonDistributed SystemsFull-Stack DevelopmentAWSDockerKubernetesMachine LearningSystems EngineeringResearch

About the role

  • Think of them as "AWS for AI models"-not data or raw compute, but a full-stack backend for fine-tuning, reinforcement learning, inference, and long-term model maintenance.
  • Their promise is simple: they make your AI system better.
  • This is an ideal role for an early-career engineer who wants to work on real distributed systems, GPU workloads, and modern ML infrastructure-not dashboards or CRUD apps.

Responsibilities

  • Cloud & Systems Engineering - Work on cloud-native systems using containers and orchestration (e.g., Kubernetes). - Optimize systems for performance, reliability, and cost efficiency, especially for GPU-heavy workloads. - Implement monitoring, logging, and observability for long-running training jobs and production services.
  • Collaborate with ML Engineers - Partner closely with ML engineers to support evolving model architectures, training workflows, and evaluation needs. - Translate ML requirements into scalable backend and infrastructure solutions.
  • Their customers are Series A-C AI companies building enterprise-grade products.

Requirements

  • Required - 1-3 years of backend engineering experience, ideally working on production systems..

Nice to have

  • Strongly Preferred.
  • Experience with or exposure to ML infrastructure or ML platforms. - Familiarity with GPU workloads, training pipelines, or inference systems..

Benefits

  • Rockstar is recruiting for a fast-growing startup that is building the AI backbone for the next generation of intelligent products.
  • They help fast-growing AI startups design, fine-tune, evaluate, deploy, and maintain specialized models across text, vision, and embeddings.
  • SGLang - or similar distributed ML systems Bonus

This listing is sourced directly from Rockstar 3's careers page and normalized into a canonical job model.