GPU Kernel Engineer for AI Inference & Performance

FriendliAI; San Francisco (CA); On-site; USD 120,000 - 150,000; Full time

Benefits offered by this job

Flexible working hours

Daily lunch and dinner

Health check-up support

Competitive compensation

Startup equity

Health insurance

Job summary

FriendliAI is seeking a GPU Kernel Engineer in San Francisco to design and optimize GPU kernels for AI inference. This role requires expertise in CUDA, C++, and performance-critical systems. You will work on cutting-edge GPU technology and contribute to a highly collaborative, supportive work environment with competitive compensation and startup equity. Join us at the forefront of AI infrastructure innovation.

Qualifications

• 3+ years of experience in GPU programming.
• Strong proficiency in CUDA for NVIDIA GPUs or ROCm/HIP for AMD GPUs.
• Deep understanding of GPU architecture and tuning.

Responsibilities

• Design, implement, and optimize high-performance GPU kernels for AI inference.
• Develop and maintain GPU code in CUDA and C++.
• Benchmark and ensure performance parity between NVIDIA and AMD hardware.

Skills

GPU programming

Performance-critical systems

CUDA proficiency

C++ proficiency

Performance tuning

Education

Bachelor’s or Master’s in Computer Science, Computer Engineering, Electrical Engineering

Tools

CUDA

C++

ROCm/HIP

Job description

FriendliAI is seeking a GPU Kernel Engineer in San Francisco to design and optimize GPU kernels for AI inference. This role requires expertise in CUDA, C++, and performance-critical systems. You will work on cutting-edge GPU technology and contribute to a highly collaborative, supportive work environment with competitive compensation and startup equity. Join us at the forefront of AI infrastructure innovation.