Senior ML Engineer - GPU-Accelerated PyTorch & GNNs

United States Digital Space LLC; United States; Remote; USD 152,000 - 288,000; Full time

Benefits offered by this job

Equity

Benefits

Job summary

United States Digital Space LLC is seeking a senior software engineer to build large-scale ML solutions including Graph Neural Networks, Tabular Foundation Models, and ensemble models with GPU-accelerated training and inference. You will accelerate PyTorch-based frameworks and integrate CUDA-X libraries to power high-performance workflows.

You will collaborate with product, engineering, and science teams to develop GPU-accelerated implementations, mentor engineers, and drive architectural

Qualifications

• Bachelor's degree or higher with 5+ years in ML, DL, and data science.
• 3+ years of PyTorch experience.
• 2+ years training enterprise-scale models on distributed infrastructure.
• 2+ years designing and operating GPU training/inference workflows, incl. profiling and scaling.
• Excellent C++ programming and software design skills.

Responsibilities

• Develop accelerated PyTorch-based solutions for large-scale models on GPU infrastructure.
• Support CUDA-X libraries and integrations in PyTorch workflows.
• Collaborate with developers, product managers, and scientists on GNNs and GPU-accelerated implementations.
• Gather customer requirements to guide product and engineering priorities.
• Provide technical leadership and mentorship across the team.
• Identify opportunities to improve the codebase and reduce maintenance.
• Apply agentic coding tools to fix bugs and refactor code.
• Solve complex issues and coordinate across teams.

Skills

PyTorch

C++ programming

GPU acceleration

CUDA

GNNs

Distributed training

High-performance computing

Education

Bachelor's degree or higher

Tools

Snowflake

Databricks

Job description

United States Digital Space LLC is seeking a senior software engineer to build large-scale ML solutions including Graph Neural Networks, Tabular Foundation Models, and ensemble models with GPU-accelerated training and inference. You will accelerate PyTorch-based frameworks and integrate CUDA-X libraries to power high-performance workflows.

You will collaborate with product, engineering, and science teams to develop GPU-accelerated implementations, mentor engineers, and drive architectural