Our client, a global trading company wanted us to develop and execute the backbone of the firm, i.e., the trading engine itself. The foundation for this platform is not just the code complexity, considering the timelines of trade, prices of bonds and the trade volumes but also the trust between brokers and potential clients who will be relying on this platform for trade execution.
Instead of building a ordinary trading platform where only buying and selling happens, we decided to take a step up to satisfy an institution with recommendations whenever there is an unavailability of the bond they wanted in the market. This recommendation is based not only on what is similar to the bond that is unavailable, but also on recent trends of buying and selling. This does not just quench the customers with diverse portfolio’s but also increases the liquidity in the market.
- The platform utilizes Dark Dutch Trading, where the central authority does not let you know what is available in the current basket of the market.
- While uploading your orders onto a platform is a heavy lift in itself, executing the orders without bias is a second huge step.
- We had integrated data from multiple institutions and we record audit trails for easy processing in data warehouses.
- Beyond the base execution and distribution, to increase liquidity in the market, we offer potential alternatives to put in your basket based on your interest. This suggestion involves collaborative filtering and user and firm buy and sell histories.
- Beyond this, instant tuning of recommendations by recording the trail using clickstream data.
TECHNIQUES, TECHNOLOGIES, TOOLS
- Java, JHipster, Python, Flask, AWS Services, Docker, Collaborative filtering
On going project.