The Machine Learning Engineer will be in charge of Lingua Custodia’s Neural Machine Translation systems in production. His/her work will be in constant collaboration with the Data Management, Research and Development, IT and product specialist teams in order to design and implement efficient Machine Learning techniques, and to build strong Neural Machine Translation models.
- Build Neural Machine Translation systems for production
- Design and run experiments to improve translation quality: optimal hyper-parameter set, best use of specific datasets, etc.
- Implement needed functionalities in Neural Machine Translation frameworks
- Implement and evaluate state-of-the-art models
- Optimize inference in production
- Contribute to open source projects
- Improve data pre-post-processing pipeline
- Participate in user feedback processing
- Master’s degree in Computer Science with a major in Machine Learning.
- Strong knowledge of Natural Language Processing
- Good knowledge of neural sequence-to-sequence architectures: Recurrent Neural Networks, Attention mechanism, Transformers.
- Experience with domain adaptation techniques in low-resource conditions.
- Experience in Linux environment: Bash scripting.
- Proficiency in Python and/or C++
- Proficiency in at least one neural framework: TensorFlow, Pytorch, MxNet, etc.
- Knowledge of at least one Machine Translation toolkit: Sockeye, Tensor2Tensor, MarianNMT, etc.
- Experience with agile software development
- Friendly startup environment in the center of Paris
- Possibility to work remotely
- Good health insurance
Applications are expected by email: firstname.lastname@example.org
Please, specify the position name in subject field.