Research Software Engineer (m/f/div), 100%
The Deep Models and Optimization group and the Safety- and Efficiency- aligned Learning group at the Max Planck Institute for Intelligent Systems are jointly seeking one research engineer (m/f/div) with strong coding skills and mathematical/statistical background to support several research projects concerning Efficient Training of Safe Deep Learning Models. The projects are taking place in collaboration with the ELLIS Institute Tübingen for a total duration of one year. This role comprises both active coding and theoretical investigations: the candidate is ideally experienced in both the practice of deep learning models and in statistics (e.g., development of new metrics and supporting their effectiveness with theory and experiments).
Your tasks
Your role will be to support the groups of Jonas Geiping and Antonio Orvieto in their research projects for the year 2024. The ideal start date is Jan 1st, 2024.
Responsibilities include
- Following a few projects in the groups closely (ideally, 1-2 per group): attending meetings and actively participating in ideas development.
- Developing codebases and necessary implementations supporting the project development.
- Benchmarking optimizers and deep learning models by collecting statistics and by coming up with new metrics.
- Providing visualizations and necessary ablations for publications.
- Supporting the paper writing and developments in the theoretical sections by checking proofs and proposing new solutions.
- Keeping an updated GitHub repository for each project, to be published after publication.
Your profile
- A Bachelor’s and Master’s degree in Engineering or Natural Sciences with a strong focus on mathematics and statistics (e.g. Mathematics, Mathematical Engineering, Computational Science and Engineering, Physics).
- Strong coding skills and knowledge of deep learning: knowledge of PyTorch and of the practices of training transformers is a strict requirement.
- At least 2 years of experience in the deep learning industry. Ideally, the candidate has worked for top research institutions at the intersection of optimization theory and deep learning.
- Experience with language modeling (application area in the Safety- and Efficiency- aligned Learning group) and biological/medical data (application area in the Deep Models and Optimization group) is a plus.
- Publications in deep learning or related subjects is a plus.
What we offer
The position will be located in Tübingen at the Max Planck Institute for Intelligent Systems. You can expect an international environment with regular opportunities for further personal and professional growth. You will work closely with scientists, other engineers, administrative staff, and IT. You will enjoy an open, multicultural, and informal working environment.
Salary and contract will be based on qualifications and previous professional experience according to the Public Service Wage Agreement (TVöD Bund) guidelines. All social benefits of public service apply. With us you receive a financial contribution to the job ticket (ÖPNV) and the possibility of mobile working.
The position is initially limited to one year.
Contact and application
To apply, please include a cover letter and your CV with contact details of references. Please upload it as one PDF file to our application portal (https://jobs.is.mpg.de/2gcan).
The position will be open until filled. Preference will be given to the applications received by Dec 01, 2023.
