Crate Friendly Software Development Workshop

In this training, emphasis is placed on experiences gained from machine learning projects, based on which it is good to draw conclusions and learn lessons for future projects. In addition, a practical part is also included in the training, during which new knowledge can be consolidated. The training is conducted on the principle that examples of previously completed Krati projects are given and problems are solved interactively.

Overview of Krat at a technological level

Krati's project management training focuses on how to understand what can be done with existing data and how to validate machine learning use cases as quickly and efficiently as possible.

Data control and preparation

The training focuses on data preparation, how to conduct an initial data check and assess whether there is enough data for the project.

Conducting a data science project

The aim of the training is to give public sector project managers an overview of data science projects and how to carry them out better. In particular, the training focuses on the PoC phase of data science projects, where uncertainty is still very high. The training is three-part.

Automatic extraction of information from Estonian text

The given training focuses on the automatic extraction of information from the Estonian text - existing tools, their application possibilities and the features of the Estonian language from the point of view of automatic processing. The training is intended for project managers. The training was conducted by Dage Särg.

About speech recognition and speech synthesis

The training focuses on speech recognition and synthesis from both a technological and a practical point of view. In the course of the training, the most important technical aspects and concepts are reviewed so that those who have completed the training can put various solutions into context and the possibilities and shortcomings of technology implementation.

About conducting a machine vision project

The training will focus on what it takes to run and implement a machine vision project, what different approaches are there, what is tagging, what is CNN, YOLO and more.
The training was conducted by Taavi Tammiste.

Crate-friendly software development 1

The purpose of the training is to provide an understanding of how a machine learning project is structured, how it differs from software development projects, and what are the prerequisites for the project to be implemented as efficiently as possible and with a high probability of success

Crate-friendly software development 2

The aim of the training is to give the participant a comprehensive understanding of speech recognition from both a technological and a practical point of view. During the training, the most important technical aspects, concepts, shortcomings and areas of application are reviewed.

Automatic extraction of information from Estonian text

The training focuses on automatically extracting information from Estonian text. The training covers, among other things, text segmentation, identifying time and name units and adding your own name units, creating lexicons and text lemmatization using the EstNLTK library.