Team Mirsis
Description
**Team Background**: Mirsis is a leading IT consultancy company in tailored banking and financial services solutions in Istanbul, Turkey. The R&D arm of the company develops cutting edge projects, spanning financial analytics with machine learning forecast and categorization models and modern web/mobile interfaces.
**Solution Proposal**: “An automatic lending machine” utilizing a machine learning based scoring model boosted with alternative sources of data, including “Resemblance” data partnered with banking institutions, to categorize the behaviors of non-clients, underserved clients and unbanked customers and forecast acceptable limits for each profile
**Business Rationale**: 43.1% urban population in Egypt, compared to 84.07% in Saudi Arabia, necessitates regional and customized solutions inevitably. In the 14th most populous country in the world, 56.36% of the population is aged between 15-54 years, and 27% is the youth population between 15-29 years and with a population growth rate between 1.94% - 2.05%, youth is the first targeted segment as MINT thrives to encompass.
Also, 62% of micro and small businesses does not have access to banking credits, and 40% of youth borrows from family and friends (9 million). Referenced statistical data shows the potential for huge market size. Detailed statistics will be used to form the market size estimates.
In terms of the speed of service, weeks-long processes are unacceptable within an online community of people and even unreachable for remote and underserved parts of the country. The business needs to be anti-red-tape and fast, with the help of digital documentation and lending decisions. The ultimate aim is a near real-time lending process.
**Technical Rationale**: Given the line between individuals and micro-businesses converges, the model should take the combined rating constituents into serious consideration. The machine learning algorithms (DNN & LSTM RNN) will be ready to find more about the disparate classifications of the profiles and then forecast the applicable credit limit amounts within the behavioral and demographical characteristics. Upon the decisions, the lending machine will be readily available to complete the transaction with an easily understandable interface that hugs the user and facilitates the transaction.
Bank-to-customer, peer-to-peer, or even jamāʿāt-to-peer or vice versa, channels are possible within the framework we will work. Our team will pivot various channels as the interface for our lending machine.
Though quite hard to cover all regulations, the critical and risky compliance issue candidates will be judged under EFSA, FRA, CBA rulesets.
**Remarks**: Let's work through the alternative data for the growth of EGBANK and MINT. Fast.
HG and TM
Team Members
-
Team Lead