EL240521P2B1 - Optimization & GPU Engineer (ASAP, Brussels, 6 months)

Lille, Nord-Pas-de-Calais-Picardie  ‐ Hybrid
This project has been archived and is not accepting more applications.
Browse open projects on our job board.

Keywords

Ajax (Programming Language) Nvidia CUDA Convex Optimization Python (Programming Language) Topology Gpu Programming

Description

Please send us your English CV directly to our projects2 mail with the project ID.

Service Description: Side-by-side with other Contingent Workers and internal staff, you will play a part in building a sustainable, green and reliable energy infrastructure that stimulates economic growth and reduces our ecological footprint.
Only together can we make the biggest impact possible! We believe in the power of diversity: we are convinced that an inclusive and diverse work environment leads to more innovation, creativity and better results.
Tasks:
• Implement QD/RL based optimization strategies for grid topologies
• Perform structured experimentation towards a reliable optimizer solution
• Implement jax/CUDA based environments for massively-parallel training
• Carrying out benchmarks to identify bottlenecks and for quantitative evaluation.
• (Optional) scientific publication of the knowledge gained.

Requirements:
• Experience: at least three years +
• Fluent in English. Additional language skills in German, Dutch or French is a plus
• Strong knowledge in Quality-Diversity/Model-based Reinforcement Learning/general non-convex optimization
• Knowledge of parallel programming patterns and massively-parallel hardware (GPUs)
• Strong knowledge on Python, Jax , CUDA ( of iets anders relevant)
• Fluency in at least one RL or QD framework (QDax, rllib, rlax, …)
• Good software-engineering knowledge, able to write maintainable and performant code
• BSc or MSc in Computer Science, engineering, informatics, or similar is a plus
• Background in electrical engineering is a plus
• Good communication skills


- Start: ASAP, approx. 6 months, extension possible
- Workload: 5 days per week
- Location: Working location will be mainly in the Brussels area (homeworking possible 2-3 days a week)

Important: There many agencies looking for the same role, but we can guarantee the lowest rate. To ensure a successful deal, we ask you not to introduce yourself to other agencies at a lower price than with us!

We look forward to receiving your application via mail!
Your team from joyIT Berlin
Start date
ASAP
Workload
100% (5 days per week)
Duration
4 months
From
joyIT Berlin
Published at
10.05.2024
Contact person:
Michael Hoehne
Project ID:
2748850
Industry
Energy Industry
Contract type
Freelance
Workplace
20 % remote
To apply to this project you must log in.
Register