Services

We value a transparent approach to consultancy engagements.

We strive to develop code according to the highest quality standards for software (PEP8), with a fully documented Python source. For most projects, we deliver a git-versioned, object-oriented library of documented modules to facilitate handover to other developers or machine learning engineers for further deployments on production or integrations.

On the same principle, we choose well-known, production-ready, open-source Python frameworks for data wrangling, data cleaning, feature engineering, visualization and Machine Learning development to avoid you get locked into any proprietary vendor:

• • PyTorch: for fast, flexible experimentation and efficient Deep Learning development
  • PyTorch-Lightning: to further speed up research and development
  • TensorFlow: developed by Google, it is a well-known library for distributed numerical computation and very large-scale applications
  • Keras: high-level Deep Learning API for neural networks
  • Scikit-Learn: when Deep Learning is not the best fit, we use the well-tested, robust framework for classical Machine Learning
  • SpaCy – Industrial-strength Natural Language Processing framework.