Daxx is looking for a Full-Stack (Java+Angular 7) Developer to join the MatchingLink team in Kyiv.
MatchingLink builds software for institutional investors and more specifically asset managers, insurance companies and pension funds. Our software facilitates the daily process of valuing the liabilities (for instance pensions to pay) versus a portfolio of assets (obligations, shares, etc). Based on this valuation, proposals to improve the portfolio are created based on optimization strategies. Improvements are generally along the lines of risk mitigation, yield increase, and currency exposure reduction.
The data used in the above scenario can be fed using Excel files or an API and the process is more or less automated with a detailed workflow.
When discussing and implementing the above solution with our clients we discovered that they have major problems organizing the data that should be fed in the system as well as structuring the reports coming out of the system. We, therefore, thought of a product that has a much more flexible data layer.
The basis of the system is what we call the ‘grabber’. It can recognize data models from excel, JSON and XML files. We want to add PDF data sources (using Amazon Textract) as well. Based on these data sources, users can them import, validate, transform and approve data as well as create new data from transformations.
On top of this data platform, we want to develop very specific modules for the asset management industry. The first two modules are the Mortgages valuation and Portfolio optimization modules.
The Mortgages module calculates and reports on the value of a Mortgage portfolio and. The calculations include different risk scenarios and the reports include textual data as well as (NVD3) graphs.
The Portfolio optimization module is intended to optimize investment portfolios based on criteria such as Interest Rate Risk, ESG (Sustainable investing), Return and geographical distribution. It starts with calculating the current value of the portfolio and then runs an AI optimization algorithm to propose possible optimizations. An interactive UI then help the end-user to choose the preferred optimization, possibly by changing the parameters or by combining optimizations.
A prototype of the DART is already developed. The technology stack is: