Staff Systems Engineer - Container Runtime & Linux Kernel

Modal; United States; Remote; USD 150,000 - 210,000; Full time

Job summary

Modal is seeking a systems engineer to advance the Inference Runtime team, building a production container runtime stack in Rust and Go. You will diagnose Linux and performance issues, shaping the architecture of Modal's container platform while collaborating with upstream maintainers like gVisor.

You will tackle fast startup, multi-GPU support, zero-copy data paths, memory-efficient snapshots, and performance tuning for Linux-based workloads across GPUs and RDMA-enabled setups.

Qualifications

• Strong Linux systems knowledge: processes, memory, filesystems, system calls, namespaces, cgroups.
• Experience building or debugging container runtimes, sandboxes, or low-level infrastructure.
• Proficiency in Rust or Go; interest in both languages.
• Experience profiling performance where memory movement and I/O are dominant.

Responsibilities

• Make container startup, checkpoint, and restore dramatically faster for large workloads.
• Build multi-GPU and accelerator-aware snapshotting with efficient memory handling.
• Design zero-copy and low-copy data paths between memory, filesystems, storage, and runtime.
• Optimize snapshot pipelines with parallel uploads and direct I/O.
• Improve container image and filesystem performance across EROFS, FUSE, and overlay FS.
• Extend sandboxed runtime to support new GPUs, drivers, and devices.
• Debug complex failures across system calls, VM, kernel behavior, and drivers.
• Roll out runtime changes across fleet with compatibility controls and observability.

Skills

Linux systems

Container runtimes

Rust

Go

Performance profiling

Kernel debugging

RDMA tech

Tools

gVisor

runsc

Linux kernel

Job description

Modal is seeking a systems engineer to advance the Inference Runtime team, building a production container runtime stack in Rust and Go. You will diagnose Linux and performance issues, shaping the architecture of Modal's container platform while collaborating with upstream maintainers like gVisor.

You will tackle fast startup, multi-GPU support, zero-copy data paths, memory-efficient snapshots, and performance tuning for Linux-based workloads across GPUs and RDMA-enabled setups.