Requirements: English
Company: WhatJobs
Region: vlaanderen, belgique , Flanders
Machine Learning Algorithm/SW Optimization Engineer
Join to apply for the Machine Learning Algorithm/SW Optimization Engineer role at imec .
What you will do
- Understand and optimize performance of deep learning libraries through deep knowledge of AI frameworks, algorithms, models, and related hardware.
- Set up AI workloads, analyze, and develop optimization techniques for existing and future compute architectures in collaboration with stakeholders.
- Explore new algorithms/optimizations for custom architectures to enable SW-HW co-design.
- Stay current with deep learning literature to implement state-of-the-art algorithms.
- Communicate progress within research projects to stakeholders and adapt as needed.
- Participate in the EuroHPC DARE project to advance Europe''s independence in high-performance computing and AI.
What we do for you
Join a leading research center in nanotechnology at imec''s headquarters in Leuven, Belgium. We foster an inclusive, multicultural, and innovative environment with opportunities for personal and professional growth through imec.academy. We offer a competitive salary with fringe benefits, recognizing your valuable contributions.
Who You Are
- PhD in Computer Science, Engineering, Mathematics, or related field.
- Proficient in parallel programming, CUDA, Python.
- Experience with Deep Learning Frameworks (PyTorch, TensorFlow, Jax).
- Experience with distributed training frameworks (Ray, Dask, PyTorch Lightning).
- Knowledge of optimization techniques like quantization, pruning, distributed training.
- Experience in software performance analysis and hardware optimization.
- Experience with AI compute architecture optimization is a plus.
- Strong communication skills and team-oriented mindset.
- Fluent in English, both spoken and written.
Note: IMEC accepts resumes only directly from candidates or through approved channels. Unsolicited resumes from agencies without prior agreement will not be accepted or compensated.
#J-18808-Ljbffr