Technologies

Universal Artificial Intelligence is an anthropomorphic system that is capable of self-development and decision-making in every sphere of life. It is not our aim to create such a system. We are developing technology capable of recursive self-learning, as well as having the ability to adapt to new information. This gives us the opportunity to analyze hundreds of thousands of options and keep the best ones. Moreover, validating models is extremely important in the financial sphere, and our AI can create validation algorithms to ensure stable functioning in the future. AI elements help create decision-making models on the stock market and complete a number tasks related to the identification of user preferences and data visualization. All of these elements are incorporated into the technologies described below.

Nowadays it’s easy to find correlations in numerous financial data sets and create overfitted algorithms. Most of them will be inevitably challenged by real market. Machine learning including artificial intelligence are powerful techniques but they should be applied very selectively. We believe that vigorous verification is the most important part in algorithmic trading and systematic asset management. We are of the opinion that’s the only way to avoid loses and disappointments in algorithmic solutions. We are specialists in multi factor verification and can help your quant team to stabilize and upgrade their models. We believe that together we can create for you really advanced and adoptive algorithmic systems capable to cope with hardly predictable future market conditions.

There are currently 3 solutions available for people who manage their own money on the stock market – trading terminals (which leave users to face Wall Street sharks on their own), handing money over to a fund (receiving the same return as all other investors) and robo advisers (which provide non-personalized financial advice). We think that modern technology allows us to address the needs of investors in a more comprehensive fashion and develop technology that will provide people with more personalized solutions. Social networks already recommend connecting with people we may know and give us suggestions for movies and music. Why can’t we also recommend relevant articles, financial instruments or even specific models/approaches to individuals? Much like there is no point in explaining what an ETF is to Wall Street veterans, there is no need to overwhelm a dentist or hockey player who decided to invest a small sum of money in the stock market with detailed statistical analysis.

There are a lot of ideas and ways to execute them on the stock market. Some people rely on financial statistics, some look for fundamental indicators that match set values, and others simply wait for a positive momentum in price. The problem is that the market currently has thousands of tools that are impossible to consolidate by hand. This is why we use Big Data search and visualization technology – we create signal systems that allow users to find and visualize what they need

Although the stock market has one of the biggest and most structured data bases, a closer look reveals a set of difficulties. There is often no problem when it comes to working with share price, but this changes when one tries to work with non-price characteristics or derivatives. By working with fundamental data of companies, analyst evaluations, options prices and other characteristics, we have created technology that allows us to find and process discrepancies, detect errors and reconstruct accurate information. This allows us to use correct data in our solutions, as well as to conduct accurate research.

All of the above-mentioned technologies are tools for completing tasks on the stock market. But this is not possible without applying stock market models. We have recreated a large number of approaches to portfolio optimization, creation of diversified portfolios from different asset classes, hedging optimization, volatility clustering, technical analysis algorithms and much more. We supplement our algorithm library every day by testing new ideas and approaches and removing inefficient strategies/algorithms (or adding them back as they become efficient again).