Kernel Driver Software Engineer
Etched · San Jose
CHECKED 8H AGO
Infrastructureonsite$150k–$275k
At a glance
Highlights
- Build kernel drivers for new ML accelerators for frontier AI inference.
- Co-design chips, racks, and software for best-in-class throughput and latency.
- Optimize high-bandwidth data paths to minimize latency and maximize throughput.
What they offer
- Medical, dental, and vision packages with generous premium coverage
- $500 per month credit for waiving medical benefits
- Housing subsidy of $2,500 per month for those living within walking distance of the office
- Relocation support for those moving to San Jose
- Wellness benefits covering fitness, mental health, and more
- Daily lunch + dinner in office
- Unlimited compute budget subject to ROI justification
Ideal candidate
Low-level systems thinker who loves hardware-software co-design and sweating reliability and performance details to power frontier AI inference.
Requirements
- Proficiency in C/C++
- Strong understanding of kernel-mode driver development and debugging
- Deep understanding of operating system internals (Linux preferred)
- Experience with hardware/software interfacing and device drivers
- Experience with memory management and synchronization in kernel environments
- Strong understanding of PCIe and other hardware interfaces
- Experience with device virtualization technologies, including SR-IOV and VFIO
- Strong understanding of kernel memory mapping, page table configuration, and IOMMU
- Familiarity with hardware-software co-design principles
- Proven ability to analyze complex technical problems and provide effective solutions
- Excellent communication and collaboration skills
- Experience with version control systems (e.g., Git)
- Experience with debugging tools (e.g., gdb, kgdb)
- remote policy
- fully in-person
Lab profile
About Etched
Building specialized AI chips and systems for frontier inference with best-in-class throughput and latency.
What they build
- Sohu AI inference chip
- AI racks and hardware systems
- Inference software stack
Known for
- Full-stack co-design of chips, racks, software and manufacturing
- Best-in-class throughput and latency for prefill and decode
- Inference-optimized architecture
Key facts
- Backed by hundreds of millions from top-tier investors