New
Project specification
Period 9/14/26 - 2/26/27
Daily rate 720 - 960€
Languages
German (Advanced)
English (Advanced)
Project expertise
Industries
Information Technology
Areas
Information Technology Product Development Research and Development
Description
- Expand and improve our deep learning frameworks (based on PyTorch) to boost efficiency and usability across different use cases
- Scale the distributed training infrastructure and tooling for faster and larger model training operations
- Design and implement data strategies for large-scale datasets, optimize storage and GPU utilization
- Debug and resolve performance bottlenecks in production ML pipelines at scale
- Contribute technical expertise to our engineering culture and share insights on ML optimization and systems design
Requirements
- MSc or PhD in Computer Science or a STEM field with a focus on machine learning and deep learning
- Strong software engineering skills in Python; confidence with modern DL frameworks (PyTorch, JAX or TensorFlow)
- Proven experience implementing custom layers, loss functions, and distributed training loops – not just library imports
- Hands-on experience training models on large-scale GPU clusters and implementing advanced parallelism strategies (NCCL, MPI)
- Familiarity with workload orchestrators (Slurm, Kubernetes, Ray) for managing complex training jobs
Not applying this time?
Get notified about similar projects matching your experience.
Frequently asked questions
The project preferably starts in September 2026 and is planned to end in February 2027 (5 months).
The project offers a daily rate of 720 - 960€ which breaks down to an hourly rate of 90 - 120€/h.
The project requires German (Advanced) as essential language. Additionally, English (Advanced) is desirable.
The project is related to the following industry: Information Technology.
The project covers the following business areas: Information Technology, Product Development, and Research and Development.
Yes! Recommend a freelancer for the project and earn 30% of FRATCH's profits every time they get placed — for the duration of that project. Simply share your invite link with a colleague to get started.
To apply for the project, click the Apply button on the project page to submit your profile for review. We will forward your resume to the client and get back to you within a few days.
Join other experts who are already part of our network
Published: