Optimize machine learning performance at Waabi through CPU and GPU profiling, memory analysis, and productive training frameworks. Support research and production systems.
This Senior ML Training Optimization Engineer focuses on identifying performance bottlenecks across model runtime and memory. You will evaluate technologies including custom CUDA kernels, quantization-aware training and inference, and compilation techniques. Collaboration with researchers and ML engineers supports optimal resource usage.
Key Responsibilities: Comprehensively profile model runtime and memory
Pinpoint performance bottlenecks across machine learning systems
Identify emerging training and inference technologies
Work with researchers on resource usage best practices
Improve tooling and dashboards for adoption
Requirements: Relevant MS, PhD, or bachelor’s degree
Four years of industry experience with a bachelor’s degree
Python, C++, or Rust coding proficiency
PyTorch or Jax deep learning framework experience
PyTorch Profiler or NVIDIA Nsight experience
Bring performance optimization, profiling, and framework expertise to Waabi’s machine learning training and inference work.