First device prototype

Dmitry confirmed the hypothesis that it is possible to try this device.

Approximately 1.5 weeks of correcting software elements, removing unnecessary functions. After about 2 weeks I already received a device, which recorded some readings of accelerometer and gyroscope on microSD.

The idea was this – if we record some readings and have a stream of data what prevents us, using artificial intelligence to analyse a large amount of data and based on it get an understanding of “Do we have any changes relative to the horse’s normal behaviour?”.

If the horse walks every day, a pattern can be detected on the large amount of data collected and it is something akin to the work of compression algorithms. When these algorithms find similar components in some sequence of data, they are labelled and labelled with certain indices. And at the expense of processing quite large volumes of data in the file – typical data are replaced by shorter sequences, data compression takes place.

The first idea was that this data stream taken from microSD should be put into a neural network (artificial intelligence) and based on BigData we should infer changes we can see in a horse’s movements.

Then the next question arises. All right, data collection is provided. Then what to do?

We need a mathematician-analyst who can process large amounts of obtained information. A person must also be able to build neural networks.

Next, having a great desire to find an expert in this area, it was decided to go on Upwork.The first experience showed that one general description of the task and the desire of job applicants overlapped so much that there were more than 1,000 applications. But how do we find the only expert? That would have taken years. ….Further on it was decided to open a competition on the Upwork platform with the condition that the first communication was done by the client.

Gilberto Galvis

It was then easier to identify the 12 leading “experts” from a large pool of candidates. From this list in the process of discussing the task 7 people have decided that this is not an appropriate project for them. Then only 5 candidates from the sample remained. As a result of close interaction and problem solving, one expert was left – Gilberto Galvis.