🛠 Tech Stack

💼 About This Role

You'll design and implement the low-level runtime stack that drives FuriosaAI's NPU hardware, from device driver interfaces to embedded firmware. Your work will maximize inference throughput and minimize latency for real-world AI workloads. This role offers the chance to optimize across the full stack — from firmware to user-space — on cutting-edge accelerator hardware.

🎯 What You'll Do

• Develop low-level runtime for DMA I/O and kernel execution scheduling
• Build asynchronous execution pipelines for NPU hardware
• Enable multi-node inference with RDMA-based communication primitives
• Develop embedded firmware on NPU's integrated ARM core

📋 Requirements

• 3+ years systems programming in Rust, C, or C++
• Bachelor's degree in Computer Science or equivalent
• Computer architecture fundamentals: memory hierarchy, DMA, interrupts
• Strong communication for cross-team alignment

✨ Nice to Have

• Experience with low-latency runtime systems or embedded firmware
• Experience with DMA engines or zero-copy data transfer
• Familiarity with CUDA low-level runtime internals

🎁 Benefits & Perks

• 🚀 Cutting-edge AI hardware impact
• 📚 Learning opportunities with senior engineers
• 🌍 Global collaboration with international teams

📨 Hiring Process

[email protected]

🚩 Heads Up

• Mismatched level: requirement says 3+ years (mid) but title implies senior potential
• Vague 'strong communication' listed as a minimum qualification