Systems/GPU Programmer

About the Role

As a systems/GPU programmer you will play a crucial role in developing new matrix multiplication kernels and algorithms that can improve inference for AI models. You will help develop new high performance tensor libraries and auto optimization tools. Working closely with the technical CEO and founder, you will co-design GPU kernels and model architecture to enhance the performance and efficiency of our AI systems. Your ability to research and stay on top of cutting edge papers will be vital in staying up-to-date with the latest advancements in AI model inference and GPU programming techniques.

Example Projects

• Implement new high performance sparse matrix multiplication kernels
• Create highly parallel GPU code graph translation and optimization systems
• Develop or extend parallel generic GPU libraries for approximate sorting, k-max, hashtables, etc
• Help design and deploy new market based resource management systems for our host daemon
• Quickly investigate and summarize options for new system architectures

You are:

• Deeply comfortable in C++, template meta-programming, CUDA, GPGPU, linux
• Capable and willing to quickly investigate and learn new frameworks and tools
• A fan of occasional remote pair programming
• Passionate about pushing the envelope in neural network performance and tooling
• Excited about the potential of AI

Benefits:

• Work from anywhere
• Generous compensation, equity, and benefits
• Join early into a fast growing, ambitious startup
• Opportunity to grow into other technical or management roles