Software Engineer, Runtime

Responsibilities

• Develop low-level runtime for DMA-based I/O operations and kernel execution scheduling
• Build and optimize asynchronous execution pipelines for NPU hardware
• Enable multi-node inference via RDMA-based data transfer and communication primitives
• Develop embedded firmware (PERT) for integrated ARM cores to manage scheduling and synchronization
• Profile and tune system-level performance from firmware to user-space to eliminate bottlenecks

Requirements

• Bachelor's degree in Computer Science or equivalent
• 3+ years of systems programming experience in Rust, C, or C++
• Solid understanding of computer architecture (memory hierarchy, cache coherency, OS, DMA, interrupts, MMIO)
• Strong communication skills for technical alignment

Preferred Qualifications

• Expertise in low-latency runtime systems or embedded firmware development
• Experience with DMA engines, scatter-gather I/O, or zero-copy mechanisms
• Experience developing firmware for ARM-based processors
• Familiarity with RDMA and high-performance networking
• Knowledge of CUDA low-level runtime internals (CUDA Graphs, stream-based execution)
• Experience with kernel-level performance optimizations (Linux kernel modules, eBPF, perf)
• Understanding of deep learning inference workloads and hardware execution characteristics