Automatic incident routing system

The system was tested at the largest Russian retailer by sorting real internal incidents into 270 classes. The system showed an accuracy rate of 85% (170000/200000 correctly identified classes of incidents), which was comparable to the accuracy of specialists and made it possible to almost completely replace human labor.

Urban Passenger Transport Application

The application monitors movement of the Moscow Land Urban Passenger Transport in real time. The application automatically calculates travel time and fare: • in the bus, the tram and the trolley • by suburban trains, Aeroexpress, MCC and Metro. The mobile application can also plan a route taking into account available parking or bicycle tracks. Tech: iOS, Android, Java, ObjectiveC Swift Android iOS

Banking: credit risk estimation system using AI

Client: Russian Standard Bank Core: prohibitively high credit verification costs prevented development of consumer credit business in Russia; a cost effective method to estimate credit risk, in the absence of credit history, was required to develop this market. NTR’s solution: custom algorithms, a neural network, and an extended questionnaire to estimate credit worthiness/risk quickly and cheaply. Key result: The bank has become one of Russia's market leaders in consumer credit.