Working at Soda
- At Soda, you will find a unique mix of leading scientific research and technical expertise to leverage machine learning and statistical signal processing for data-science applications that matter. Working at Soda is a unique opportunity to develop your expertise.
- Soda is part of Inria, the French computer science institute. This gives us access to collaborations with experts in high-performance computing, databases, machine learning etc.., support services with expert programmers, grid computing and massive storage infrastructures. On top of that, having worked at Inria is a great line on a CV.
- Soda has close historical ties with the Parietal and scikit-learn team at Inria: we share office space, some of the meetings, and know each other very well. As a result, we form a vibrant human, technical, and scientific community.
- Soda is committed to high-quality work, both in terms of publications and of open-source software. At Soda you can have a real impact by joining a team that produces methods and software to tackle difficult data-science problems.
- We value human interactions, all forms of diversity and personal interests.
Masters-2 level internship, typically leading to a PhD on a similar topic:
- Large-scale embedding of heterogeneous information (data science, deep learning, knowledge-graph embeddings), with Gaël Varoquaux & Fabian Suchanek
- Prediction with missing values (role of multiple imputation, uncertainty, MNAR), with Marine le Morvan & Gaël Varoquaux
- Optimizing Human Learning (reinforcement learning, partially-observable Markov decision processes, educational data mining, psychometrics) with Jill-Jênn Vie
Neural methods for relational data: We are looking for a post-doc to work on database embeddings for data science applications.
Collaboration on a hospital database: Ingénieur(e) data science collaboration Inria – AP-HP Entrepôt de données de Santé (EDS)
Machine learning programmer: We are looking for a software engineer interested in developing a plugin system to allow for efficient GPU computing kernels for popular machine learning algorithms in scikit-learn (nearest neighbors, k-means, T-SNE…).