Software Engineer III (ML Engineer)

Job Description: Software Engineer III (ML Engineer)

Location: Redmond, WA

Job Summary

We are seeking a Software Engineer III (ML Engineer) to support a research project focused on maintaining, refactoring, and optimizing machine learning models. The ideal candidate will have strong hands-on experience with Python, PyTorch, and GPU optimization, and will collaborate closely with research scientists on model training, experimentation, and evaluation.

Key Responsibilities

• Refactor, maintain, and optimize existing machine learning models and codebases.
• Improve model training and inference performance, particularly on GPU-based systems.
• Develop, maintain, and debug ML workflows using Python and PyTorch.
• Collaborate with research scientists to support model training, experimentation, evaluation, and performance analysis.
• Evaluate model behavior and performance using appropriate metrics and testing approaches.
• Identify and resolve bottlenecks related to GPU utilization, memory usage, computational efficiency, and training speed.
• Clean up, organize, and improve existing ML code to enhance readability, maintainability, and reproducibility.
• Support the integration of research models into reliable, maintainable code and workflows.
• Document model configurations, optimization techniques, experiments, and evaluation results.

Required Qualifications

• Strong programming skills in Python.
• Hands-on experience with PyTorch and machine learning models.
• Practical experience optimizing ML workloads on GPUs.
• Understanding of model training, evaluation, debugging, and performance optimization.
• Ability to read, understand, refactor, and maintain existing ML code.
• Strong analytical and problem-solving skills.
• Ability to collaborate effectively with researchers and engineering team members.

Preferred Qualifications

• Experience working directly with research scientists or ML researchers.
• Experience cleaning up, restructuring, and organizing research-oriented ML codebases.
• Familiarity with GPU profiling, memory optimization, and efficient model training techniques.
• Experience improving the reproducibility and maintainability of ML experiments.