Software Engineer, AI accelerator Runtime

📋 Description

• Design and implement the low-level device runtime for OpenAI custom silicon.
• Build kernel-launch scheduling, command submission, queueing, dependency tracking, and completion
• Manage device memory spaces, allocation, virtual-to-physical mappings, data movement, and lifetime
• Implement synchronization primitives, events, barriers, streams, and ordering guarantees that are
• Define clean interfaces between the runtime, drivers, firmware, compiler-generated code, kernels
• Use event-based, cycle-accurate simulators to develop, validate, debug, and performance-tune

🎯 Requirements

• Strong low-level systems programming experience in C, C++, Rust, or comparable environments.
• Experience building runtimes, drivers, firmware, OS components, accelerator software, or related
• Understanding of concurrency, synchronization, asynchronous execution, queues, events, and
• Knowledge of memory management, address spaces, DMA, caching, coherency, and hardware-software
• Experience with event-based, cycle-accurate simulators or related architectural/performance models.
• Ability to reason about scheduling, latency, throughput, utilization, contention, and resource

🎁 Benefits

• OpenAI is an equal opportunity employer and values diverse perspectives.
• Compliance with applicable equal employment opportunity policies and privacy practices.