Staff AI Kernel Engineer

Job Title

Staff AI Kernel Engineer

Role Summary

As a Staff AI Kernel Engineer on the SDK team, you will design and implement high-performance AI/LLM kernels and operators for Quadric’s General-Purpose Neural Processing Unit (GPNPU). You will build and mature the core kernel library, profile performance across hardware variants, and collaborate with compiler and systems engineers to maximize hardware utilization.

This role is based in the Pune office and focuses on production-grade kernel development, optimization, and benchmark-driven performance improvements for edge inference workloads.

Experience Level

Senior - requires 5+ years of experience in AI kernel development and performance optimization.

Responsibilities

Primary responsibilities include kernel development, profiling, and platform integration.

• Develop efficient AI and LLM kernels/operators in C/C++ and Python for the Quadric GPNPU.
• Build and maintain a production-grade kernel library supporting modern AI architectures.
• Profile and analyze kernel performance: compute throughput, data movement, and parallel execution.
• Optimize kernels across hardware configurations to maximize utilization and eliminate bottlenecks.
• Identify micro-architecture and software bottlenecks and implement high-performance solutions.
• Establish benchmark-proven performance standards and optimization patterns for SDK customers.
• Collaborate with compiler, tooling, and systems engineers to integrate kernels into the SDK.

Requirements

Must-have and preferred skills for successful candidates.

• 5+ years of hands-on experience in AI kernel development and performance optimization.
• Proficiency in C/C++ and Python.
• Direct compute development experience with at least one of: CUDA, DSP, NEON, or Triton-lang.
• Proven track record in model and kernel inference performance profiling.
• Strong structured problem-solving, debugging, and technical communication skills.
• Preferred: assembly language programming and experience in target NPU or semiconductor environments.

Education Requirements

Bachelor's or Master's degree in Computer Science, Electrical Engineering, or a related technical field.

About the Company

Company: Quadric

Headquarters: Burlingame, California, United States

Quadric is building the world’s first supercomputer designed for the real-time needs of edge devices. Founded in 2016, the company empowers developers across industries with innovative general-purpose neural processing unit (GPNPU) architecture for neural network workloads. Co-founded by technologists from MIT and Carnegie Mellon, Quadric aims to enable groundbreaking technology development.