Founding ML Inference Performance Engineer

uRun; San Francisco (CA); On-site; USD 120,000 - 160,000; Full time

Benefits offered by this job

Health, dental, and vision

401(k) participation

Flexible spending accounts

Paid time off

Access to AI tools

MacBook Pro and AirPods

Job summary

uRun, located in San Francisco, is seeking a founding ML Performance Engineer to drive AI infrastructure performance. In this role, you will write custom CUDA kernels and optimize model inference for real-time applications, significantly impacting performance across the stack.

The ideal candidate will possess deep knowledge of CUDA, experience with AI workloads, and a strong capacity for optimization. The position offers a competitive salary, equity, and top-tier tools for an exceptional contributor.

Qualifications

• Deep, hands-on CUDA expertise with production kernels.
• Strong background in model inference at scale.
• Fluency in hardware-aware algorithm design.

Responsibilities

• Write custom CUDA kernels for performance improvements.
• Optimize model inference targeting sub-50ms latency.
• Profile and benchmark full inference pipeline.

Skills

CUDA expertise

Model inference optimization

GPU memory hierarchy knowledge

Benchmarking and profiling

Tools

PyTorch

TensorRT

Job description

uRun, located in San Francisco, is seeking a founding ML Performance Engineer to drive AI infrastructure performance. In this role, you will write custom CUDA kernels and optimize model inference for real-time applications, significantly impacting performance across the stack.

The ideal candidate will possess deep knowledge of CUDA, experience with AI workloads, and a strong capacity for optimization. The position offers a competitive salary, equity, and top-tier tools for an exceptional contributor.