Development of a microservice with API REST in Symfony 2 and a control panel. Implemented in Symfony, Twig, and Doctrine; including database transactions management, and unit and functional testing with PHPUnit and Selenium, all deployed in a Galera cluster over Docker.
We have worked for:
Analysis, design, development, and implementation of an online translation management system, including translation and revision workflow. Integration with their machine translation engines (Politraductor) and authentication services (RedIRIS), and implemented in NodeJS, AngularJS, Handlebars, and Oracle SQL.
Integration of a customized web crawler to automate the workflow derived from website translations, and development of an automatic system for website translation budget estimation. The work was implemented in NodeJS, AngularJS, Redis, and MongoDB.
Estimation of the pose of all people in a retail establishment, with the purpose of understanding how customers spend their time there. We reached high-precision results thanks to leveraging deep learning techniques.
We developed a recommendation engine with deep learning on top of MXNet, packaged as a microservice leveraging Docker. The engine is currently receiving a rate of 25 hits/second.
Starting with models from OpenNMT (neural machine translation), our team worked to reduce the model sizes to ensure that they fit into a mobile device for offline translation.
We built a digital assistant (Alexa-like) using advanced neural networks for intent parsing, with Android and web user interfaces.