Service for online stores: analyzes the image uploaded by the user and offers similar products. Works with pictures, screenshots, photographs. Improves the performance of recommender systems.
Service for automatic broadcasting of sports events. It is based on a neural network that replaces the broadcast director. The system monitors the ball and the players, selects the best plans and switches between cameras.
The effect for mobile videos: the person and the background are painted in different colors, a colorful animation appears. To create the effect, we used an ensemble of neural networks.
The neural network monitors the growth of plants in greenhouses: evaluates their dimensions, calculates the approximate weight and "reports" when there are deviations from the norm. The service is integrated with Telegram: uploaded a photo of a plant - received an assessment
An application that takes meter readings from a photograph. Thanks to computer vision technology, it recognizes numbers in a picture better than an ordinary person.
An algorithm that evaluates the success of players in popular shooters, taking into account the difficulty level of each game.
SkinClub
An algorithm that estimates the probability of receiving a particular gift in a computer game. Supports the excitement in the players and does not allow the owner of the game to go into the red.
We found the root causes of failed transactions in the acquiring system, identified groups of problematic providers, and tested the customer's failure hypotheses.
Extensive experience in ML projects implementation
Own ML tools
Community of ML professionals
The EORA team are multiple winners of the Kaggle Data Scientist Competition. Our employees teach machine learning courses at SkillFactory, Skillbox, Innopolis University, speak at AI conferences.
In development, EORA uses its own environment for training and testing computer vision models, the InferOxy computation balancing system, and the VisionHub platform for hosting ready-made models. Thanks to this, the work on projects is faster, and the products themselves are of higher quality.
EORA is the core of the AI Community Innopolis machine learning community. Now there are over 800 people in it. Thanks to AIC, we are constantly expanding our expertise in machine learning and can quickly find the right people for a large and complex project.
02
03
01
Need help preparing a project? Order ML audit
We will help you describe the task, select the optimal metrics for evaluating the project, correctly organize the work on it, calculate the estimate and prepare the terms of reference.
EORA VisionHub maintains stable operation of ML-models and automatically scales if the load on the service increases. The EORA team helps with the preparation of the API for integrating the model into third-party services and provides its owner with access to statistics about its operation.