Distributed ML Systems Engineer: Scale Training & Tools

SIG; Philadelphia (Philadelphia County); On-site; USD 140,000 - 210,000; Full time

Job summary

Susquehanna is seeking a Machine Learning Systems Engineer to enhance distributed training infrastructure, enabling researchers to run large-scale experiments efficiently. You will work with researchers to optimize hardware utilization and build tooling for distributed workloads.

The role emphasizes Python software engineering, systems-level optimization, and profiling to improve performance across distributed training environments.

Qualifications

• Experience with large-scale ML training pipelines and distributed training frameworks.
• Strong software engineering skills in Python.
• Passion for diving deep into systems implementations and understanding fundamentals to improve their performance and maintainability.
• Experience improving resource efficiency across distributed computing environments by leveraging profiling, benchmarking, and implementing system-level optimizations.

Responsibilities

• Collaborate with researchers to enable them to develop systems-efficient models and architectures
• Apply the latest techniques to our internal training runs to achieve impressive hardware efficiency for our training runs
• Create tooling to help researchers distribute their training jobs more effectively
• Profile and optimize our training runs

Skills

ML pipelines

Python programming

Distributed systems

Performance optimization

Benchmarking

Resource profiling

Tools

Python

Distributed training

Profiling

Benchmarking

System optimization

Job description

Susquehanna is seeking a Machine Learning Systems Engineer to enhance distributed training infrastructure, enabling researchers to run large-scale experiments efficiently. You will work with researchers to optimize hardware utilization and build tooling for distributed workloads.

The role emphasizes Python software engineering, systems-level optimization, and profiling to improve performance across distributed training environments.