Principal Software Engineer, Agent Infrastructure
Core
Building and operating infrastructure for autonomous AI agents at Roblox, ensuring safety, reliability, and scalability for millions of creators.
Role type
Principal Software Engineer, Agent Infrastructure
Builds
Roblox's core agent infrastructure serving internal and external customers
Domain
AI Agents, Distributed Systems, Infrastructure Security
Deliverable
production ML models
Required skills
Backend system design, distributed systems, workflow orchestration, agentic runtimes, secure sandboxing, Go/Python/Rust, Kubernetes, observability, SLO definition, team mentorship
Preferred skills
gVisor, Firecracker, microVMs, Temporal, AWS, self-hosted models, GitOps, Argo CD, open source contributions
Responsibilities
Design and implement core agent infrastructure, maintain long-running agent reliability during deploys and outages, define security boundaries for untrusted code, set SLOs and observability standards, mentor senior engineers, align with Security and Infrastructure teams
Seniority
Principal, hands-on IC with strategic direction
