Self learning credit ratings

We have a specialist design and development team with experience of developing algorithms for automated loan underwriting, forecasting and credit scoring systems.

One system that we developed for a Bank was trained to calculate the maximum loan value, based on personal background, credit information and repayment history with minimum risk to the Bank. Another supports a loan manager during client calls, using emotional states detected in the voice signal.

Part of the Ocutri team worked on projects for Boeing to build a system for abnormal behaviour detection and forecasting.

We have built a system that automatically calculates ratings of the participants in the financial market (banks, mutual funds, etc.) and a big data system for profiling Wi-Fi users in the Moscow Metro. The software analyses their behaviour on the Internet, reconstructing gender, age, interests, income level, places of interest and other traits.

We are working with Neurodata Lab to provide systems for remote visual detection of emotions and gaze behaviour to discover emotional states and people’s intentions.

We work with both structured and unstructured, indirect or seemingly unrelated data.

Contact Us


We have years of experience in the most popular Java and JS technologies and tools for web and mobile software, such as:

  • – Meteor, Play Framework, Reactor and Eclipse Vert.x, Spring Framework

  • – Client Side: Bootstrap, AngularJS and Vue.js

  • – Big data processing. We use Cloudera CDH and customize it with other components of Apache big data ecosystem. We use Hadoop, Spark, Storm, Kafka, Flume and other components.

  • – SQL, NoSQL and Key-value databases, such as Redis, MongoDB, Cloudera Impala and Apache Hive, PostgreSQL, MS SQL.

For high load bottlenecks and computation-expensive tasks we write components in C++, C and CUDA.


Our team members have taken part in many types of financial services software systems. That's why we can offer you our help in the following areas:

  • – Automated Loan Underwriting
  • – Credit Scoring
  • – Churn rate prediction
  • – Big Data Analysis
  • – And so much more

We use machine learning, iterative and direct mathematical methods for optimisation, forecasting, finding latent variables in your data that have the most influence, evaluating risks and finding abnormal behaviour. We also write systems that do this automatically in your production environment.

You get plain numbers, meaning that you can directly control the efficiency of your business, make decisions and measure results.

We use Python, Matlab and R ecosystems for scientific R&D.

Contact Us

Professional R&D