5 days ago
Zürich, SwitzerlandMid Level / Senior
Responsibilities
- Convert research prototypes into tested, reusable data generation, training, and evaluation pipelines.
- Build distributed experiment support for data loading, checkpointing, and collecting agent interactions.
- Profile workloads and improve GPU utilization, memory efficiency, and data throughput.
- Develop tests, experiment tracking, and debugging tools while preserving experimental result integrity.
- Collaborate with researchers to turn ideas into robust, maintainable software.
Requirements
- Typically 2 to 10 or more years of relevant experience building machine learning training, inference, or data pipelines in Python.
- Strong Python skills and hands-on experience with PyTorch, JAX, or a comparable machine learning framework.
- Experience building and debugging software for model training, inference, or data processing at scale.
- Understanding of how data, numerical precision, and implementation choices affect machine learning experiment results.
- Sound software engineering practices including testing, profiling, version control, and documentation.
- Experience with PyTorch distributed, DeepSpeed, or Ray is helpful.
- Familiarity with Transformers libraries, vLLM, or experiment tracking tools is helpful.
- Background in reinforcement learning systems or working with video, audio, or image datasets is a plus.
Benefits
- Competitive compensation with equity.
- Health, dental, and vision coverage.
- Professional development support.
- Visa sponsorship and relocation support are available.
- Full-time, on-site role in Zürich, Switzerland.