Remote Platform Architect for Safe AI Runtimes

Superserve (YC F26); San Francisco (CA); Hybrid; USD 165,000 - 230,000; Full time

Job summary

Superserve (YC F26) in San Francisco / Remote is seeking a Software Engineer, Platform (Senior/Staff) to scale our runtime infrastructure and own the platform end-to-end from low‑level runtime design to production reliability.

You’ll design an agent runtime, enforce security for untrusted code, and develop scheduling/orchestration systems with Kubernetes‑level scope without copying Kubernetes, across regions and bare metal.

Qualifications

• Experience designing and operating distributed systems at scale.
• Strong systems background including concurrency, isolation, networking, and performance.
• Proficient in Go and/or Rust or willingness to learn quickly.
• Deep intuition for security and reliability, beyond feature work.
• Ability to reason about failures and system behavior under load.
• Have taken systems from 0→1 or 1→scale.
• Comfortable making high-stakes architectural decisions with incomplete information.

Responsibilities

• Design and build our agent runtime infrastructure (Firecracker, Rust, Go).
• Define and enforce security boundaries for running untrusted code.
• Architect and implement scheduling/orchestration systems (Kubernetes-level problems, without blindly copying Kubernetes).
• Implement a global scale distributed systems architecture spanning across regions and bare metal server providers.
• Own performance, isolation, and reliability at scale.
• Set technical direction and raise the bar for engineering across the company.

Skills

Distributed systems

Security & reliability

Go and/or Rust

Architectural decision-making

Scalability & performance

Tools

Kubernetes

Firecracker

Rust

Go

Containers

Job description

Superserve (YC F26) in San Francisco / Remote is seeking a Software Engineer, Platform (Senior/Staff) to scale our runtime infrastructure and own the platform end-to-end from low‑level runtime design to production reliability.

You’ll design an agent runtime, enforce security for untrusted code, and develop scheduling/orchestration systems with Kubernetes‑level scope without copying Kubernetes, across regions and bare metal.