Engine Architect for Long-Running AI Agent Runtime

Runta; San Mateo (CA); On-site; USD 180,000 - 260,000; Full time

Job summary

Runta is building runtime infrastructure to execute AI agents reliably over long horizons. This role focuses on the execution engine, including kernel and VMM work, memory layout, and crash recovery.

You will enable snapshot, resume, and branch capabilities with byte-for-byte fidelity and auditability across heterogeneous hardware. Ideal candidates have deep experience with kernels, hypervisors, or language runtimes, and are comfortable with Rust, C, or C++.

Qualifications

• Built systems at the engine level (kernel, VMM, or runtime).
• Strong memory layout, syscall boundaries, and crash recovery intuition.
• Ability to reason about copy-on-write and page sharing from first principles.
• Comfort with Rust, C, or C++ at a low level.

Responsibilities

• Design and build snapshot/resume/branch mechanics for long-running agent processes.
• Build crash recovery for stateful execution units.
• Own the microVM execution layer and optimize boot/restore times.
• Make execution recordable and auditable.
• Work at the boundary between runtime and isolation layers to manage persistence.
• Set the technical bar for the systems team as an early engineer.

Skills

Kernel/engine development

Rust

C/C++

Virtualization (KVM/Firecracker)

Snapshot/Crash recovery

Memory management

Tools

CRIU

io_uring

KVM

Job description

Runta is building runtime infrastructure to execute AI agents reliably over long horizons. This role focuses on the execution engine, including kernel and VMM work, memory layout, and crash recovery.

You will enable snapshot, resume, and branch capabilities with byte-for-byte fidelity and auditability across heterogeneous hardware. Ideal candidates have deep experience with kernels, hypervisors, or language runtimes, and are comfortable with Rust, C, or C++.