Remote GPU Performance Engineer for Large-Scale Models

United States Digital Space LLC; United States; Remote; USD 140,000 - 210,000; Full time

Benefits offered by this job

Five weeks paid leave

Comprehensive healthcare (vision +</p>

Job summary

United States Digital Space LLC seeks an experienced GPU Performance Engineer to design and implement improvements to our training infrastructure and directly influence decisions that optimize model performance.

You will contribute to post-training processes, including reinforcement learning and fine-tuning, and help scale GPU workloads across large compute clusters while shaping our serving infrastructure.

Qualifications

• Strong engineering background with Python and PyTorch.
• Experience training large deep learning models.
• Experience writing and debugging GPU code (CUDA, C++).
• Experience scaling GPU workloads with compute clusters (Slurm/Kubernetes).
• Ability to analyze and optimize GPU-accelerated performance.

Responsibilities

• Design and implement improvements to training infrastructure.
• Contribute to technical decisions to optimize model performance.
• Work on post-training processes including RL and fine-tuning.
• Improve efficiency and scalability of model serving infrastructure.

Skills

Python

PyTorch

Large DL models

GPU profiling

Performance tuning

CUDA

C++

Slurm

Kubernetes

Tools

CUDA

Kubernetes

Slurm

Job description

United States Digital Space LLC seeks an experienced GPU Performance Engineer to design and implement improvements to our training infrastructure and directly influence decisions that optimize model performance.

You will contribute to post-training processes, including reinforcement learning and fine-tuning, and help scale GPU workloads across large compute clusters while shaping our serving infrastructure.