Real-Time Vision Engineer (CUDA/PyTorch)

Moondream AI; San Francisco (CA); On-site; USD 180,000 - 250,000; Full time

Benefits offered by this job

Health insurance

Dental insurance

Vision insurance

Paid parental leave

Relocation support

Job summary

Moondream AI in San Francisco is seeking a Research Engineer to design architectures, train models end-to-end, and optimize CUDA kernels for real-time video applications. The role blends research and engineering, with hands-on execution and ownership of the model lifecycle.

You will work on Python, PyTorch, and CUDA, striving for both accuracy and speed while delivering deployable results, with meaningful equity and in-person collaboration.

Qualifications

• Proven experience training models from scratch.
• Ability to implement architectures and kernels.
• Read papers and implement results quickly.

Responsibilities

• Design model architectures and train them end-to-end.
• Write kernels and optimize runtime.
• Analyze failures and iterate on experiments.

Skills

Python

PyTorch

CUDA

Tools

CUDA

Job description

Moondream AI in San Francisco is seeking a Research Engineer to design architectures, train models end-to-end, and optimize CUDA kernels for real-time video applications. The role blends research and engineering, with hands-on execution and ownership of the model lifecycle.

You will work on Python, PyTorch, and CUDA, striving for both accuracy and speed while delivering deployable results, with meaningful equity and in-person collaboration.