Applied ML Engineer—Production Diffusion Pipelines

Drafted; San Francisco (CA); On-site; USD 150,000 - 210,000; Full time

Job summary

Drafted in San Francisco is seeking an ML engineer to help build core infrastructure for training and running our diffusion-model based systems in production. You will collaborate with engineering and research teams to deliver reliable, scalable ML systems.

The role emphasizes production-grade PyTorch code, multi-node training, and potential work with Triton/CUDA kernels to optimize performance. This is a hands-on, in-person position in SF with a small, driven team.

Qualifications

• Experience writing production serving code for diffusion models.
• Experience writing multi-node training code for diffusion models.
• Bonus: experience with custom Triton or CUDA kernels.

Responsibilities

• Build core infrastructure for training and serving ML models in production.

Skills

Production serving

Multi-node training

Model distillation

Tools

PyTorch

Triton

CUDA kernels

Job description

Drafted in San Francisco is seeking an ML engineer to help build core infrastructure for training and running our diffusion-model based systems in production. You will collaborate with engineering and research teams to deliver reliable, scalable ML systems.

The role emphasizes production-grade PyTorch code, multi-node training, and potential work with Triton/CUDA kernels to optimize performance. This is a hands-on, in-person position in SF with a small, driven team.