Senior Systems Engineer, Inference Runtime

Fdetimes; San Francisco (CA); Vor Ort; USD 180.000 - 280.000; Vollzeit

Zusammenfassung

Modal is seeking a systems engineer to own production runtime infrastructure in Rust and Go, advancing the container runtime stack used for inference and training workloads. You will diagnose Linux performance issues, shape architecture, and collaborate with gVisor maintainers to push fixes upstream.

You will contribute to running workloads across single GPUs to large multi-GPU setups, ensuring secure isolation, fast startups, and scalable performance.

Qualifikationen

• Proven strong Linux systems knowledge, including processes, memory, filesystems and signals.
• Experience building or debugging container runtimes and sandboxes.
• Proficient in Rust or Go; willingness to grow in both languages.
• Ability to profile and optimize memory movement, I/O and kernel interactions.

Aufgaben

• Make container startup, checkpoint, and restore dramatically faster for large workloads.
• Build multi-GPU and accelerator-aware snapshotting and memory handling.
• Design zero-copy and low-copy data paths between container memory, storage, and runtime.
• Optimize large snapshot pipelines with parallel uploads, direct I/O, and memory optimizations.
• Improve container image and filesystem performance across EROFS, FUSE, overlays, and remote storage.
• Extend sandboxed runtime to support new GPUs, drivers, profiling tools, and device capabilities.
• Debug failures across system calls, virtual memory, process lifecycle, kernel behavior, drivers, and isolation.
• Roll out runtime and kernel changes with compatibility checks, scheduling constraints and observability.

Kenntnisse

Linux systems

Rust

Go

Kernel concepts

Performance debugging

Container know-how

Tools

gVisor

OCI runtimes

RDMA

Direct I/O

Jobbeschreibung

Modal is seeking a systems engineer to own production runtime infrastructure in Rust and Go, advancing the container runtime stack used for inference and training workloads. You will diagnose Linux performance issues, shape architecture, and collaborate with gVisor maintainers to push fixes upstream.

You will contribute to running workloads across single GPUs to large multi-GPU setups, ensuring secure isolation, fast startups, and scalable performance.