Software Engineer, Kernel Programming

Job Title

Software Engineer, Kernel Programming

Role Summary

Implement and optimize low-level kernels for FuriosaAI's kernel programming stack to enable high-performance inference on the company’s AI accelerator architectures. Work across software, research, and product teams to integrate VLA, vision, and multimodal models and to produce developer-ready reference implementations.

Experience Level

Mid-level. No specific years-of-experience listed.

Responsibilities

Primary responsibilities focus on kernel development, performance optimization, tooling, and developer enablement for AI models on FuriosaAI hardware.

• Design and implement efficient kernels using the kernel programming stack (including vISA, TCL) targeting Tensor Contract Processor (TCP) architectures.
• Profile and optimize kernel performance using profiling tools and roofline analysis for RNGD-accelerated AI models.
• Develop automated kernel generation and optimization pipelines for AI workloads.
• Build diagnostic tools and testbeds for robust kernel validation.
• Create reproducible guides, reference implementations, and end-to-end programming enablement materials for RNGDs.

Requirements

Must-have technical skills and collaboration experience; preferred items listed separately.

• Must-have: Experience in low-level systems programming targeting XPU architectures (e.g., NPU, GPU).
• Must-have: Demonstrated experience diagnosing and improving kernel performance (profiling, optimization).
• Must-have: Experience collaborating across engineering, research, and product teams to align software with product requirements.
• Nice-to-have: Experience optimizing high-performance kernels on AI accelerators (GPU, TPU) for production AI products.
• Nice-to-have: Understanding of XPU architecture, data movement, and software–hardware co-optimization strategies.
• Nice-to-have: Experience with open-source or research projects involving model architectures such as Diffusion, Mamba, or VLA, and designing efficient deep learning architectures.

Education Requirements

Minimum: BS in Computer Science, Artificial Intelligence, Electrical Engineering, or a related field. Preferred: MS or PhD in Computer Science, Artificial Intelligence, Electrical Engineering, or a related field. The posting allows related technical fields; no explicit "or equivalent experience" language was provided.

About the Company

Company: FuriosaAI

Headquarters: Seoul, South Korea

FuriosaAI develops high-performance, energy-efficient AI inference hardware and software. Founded in 2017 by semiconductor and AI engineers, the company builds AI-native compute platforms to reduce AI energy and operational costs and operates globally with offices in Korea, Silicon Valley, and an R&D lab in Lisbon.