The results of the first EKF hackathon at the intersection of engineering technologies and machine learning on the Codenrock platform have been announced. The EKF AI Challenge brought together the best experts in the field of Data Science. The main goal of the hackathon was to create an automated service for reading electrical single-line diagrams and recognizing conductors, switches, protective devices, sensors, and other elements from the provided nomenclature database and generating project estimates.

A few figures from the competition:

·         440 participants registered.

·         210 teams worked on the task.

·         23 solutions uploaded.

·         15 projects made it to the final, their authors gave presentations of their projects to experts from EKF.

·         3 winners shared a prize fund of 500,000 rubles.

The most promising solutions for the automatic estimate calculation system were presented by:

·         1st place – ankhafizov team, 250,000 rubles. The winner developed and presented a software package with built-in recommendation artificial intelligence to create estimates with a list of articles based on electronic diagrams. The main focus was on obtaining an interpretable result from the neural network module with the ability to adjust this result.

·         2nd place – AI integration team, 150,000 rubles. The participants, who took second place, developed a product for automating the analysis of electrical circuits and providing a report via a Telegram bot.

·         3rd place – the "Almost data scientists" team, 100,000 rubles. The solution, which took third place, allows you to upload images in PDF, JPG and PNG formats, as well as manually add elements that the neural network could not recognize automatically.

Congratulations to the winners, we thank all participants for their fresh ideas and interesting projects. The best solutions from the hackathon will be implemented in the real EKF business and will help to make the best offer to customers based on the neural network.