Predictive Analytics

  • Full business process data exploration for retail, distribution and other horizontals
  • Model architecture proposal and in-depth business level presentation
  • Self-explainable model proposals in various languages and environments (Python, R)
  • Deep Learning models developed in TensorFlow
  • GPU based machine learning model training, tuning and inference
  • Predictive models operationalization in various environments
  • Advanced visualizations and User Experience

 

 

Deep Learning Research

  • Latest architectures in image localization and segmentation
  • Advanced sequence processing systems based on LSTM architectures
  • Special advanced Deep Neural Networks based on Directed Acyclic Graph architecture
  • Real time TensorFlow-based systems deployed in embedded architectures

 

 

Customer Data Science projects

  • Specifications based analysis engineering
  • Systems architecture based on customer required resources and libraries targeting various platforms (TensorFlow, Microsoft R, Python SciKit Learn, Apache Spark ML, R Caret)
  • Internally developed Machine Learning libraries (C++ and Eigen) implementing various shallow models and fully connected neural networks
  • Solution deployment in target customer environment, cloud based environment (MS Azure, etc) or fully integrated solution deployment (including high-performance computing equipment tailored for customer project needs)