Whisk is fast-growing company with the objective to "understand" food (recipes, products) and user preferences in order to provide intelligent insights and recommendations. Over several years we have been building knowledge and technology in that area, with great help of Machine Learning.
We are globally distributed and our team lives across the world in different locations – we all work together remotely. Living in a place that makes us happy is important to us.
We recently joined Samsung NEXT - the innovation division of Samsung, which allows us to accelerate our development and hire new talented people
We look for talented and experienced engineer join our team and contribute to range of products including B2C, B2B and internal applications. You will need to make sure that result application meet high-quality standards.
What you’ll do:
Find and apply the suitable NN architectures to sequence labeling and text classification tasks. Evaluate them against current models.
Working on common vector representations of Whisk core entities (recipes, users, products) to use across applications
Building recommender systems
Build models working across many languages
Minimum of 2 years working in Data Scientist role
You have good expertise in NLP domain (NER, multiclass/multilabel classification) and practical experience with NLP-related architectures (BoW, tf-idf, convolutional/recurrent NN, CRF, fasttext, transformers, etc.)
You have practical experience writing code and comfortable with Python/Jupyter environments
You are very familiar with data and visualisation tools (Pandas, NumPy, sklearn, scipy, matplotlib/seaborn)
You worked with one of popular Deep Learning frameworks, ideally PyTorch, but also good if TensorFlow
Good spoken level of English
Experience with cross-lingual transfer learning models
Great data and results visualisation skills
Experience building recommender systems
What we offer:
Please, make sure your profile information (Moikrug or LinkedIn) is up to date and a cover letter is appreciated