Intelligent preprocessing routing to decisioning services
The system and methods described herein allow users to give their applicant information when seeking to purchase a good with financing from multiple potential lenders, and may be pre-screened by one or more rule sets implemented by a marketplace client in an eligibility analysis to ultimately submit applicant information to a subset of the multiple potential lenders which are found to be suitable for lending to an applicant based on the applicant information, wherein lender microservices are then run in a jailed, firewalled, and self-contained, autonomous environment, and the results of said lender microservices are reported to the user and may be used to change the one or more rule sets implemented by the marketplace client for future pre-screening of applicants.
1. A method comprising:
encrypting, by one or more computing devices, a plurality of rule sets for prequalifying an applicant to be stored within a lender repository of a jailed environment on a server, wherein a respective rule set of the plurality of rule sets is output from each trained machine-learning model of a plurality of trained machine-learning models, wherein each trained machine-learning model of the plurality of trained machine-learning models is trained to generate rules for a respective lender of a plurality of lenders, wherein, based on the jailed environment, the respective rule set output by each trained machine-learning model of the plurality of trained machine-learning models is accessible to only the respective lender;
receiving a request to generate a prequalification result for a product based on applicant information;
for each lender of the plurality of lenders and responsive to the request to generate the prequalification result for the product based on the applicant information:
decrypting, within the jailed environment and based on a respective lender-specific key, the respective rule set,
generating, in parallel within the jailed environment, based on the respective rule set and the applicant information, a respective prequalification result, and
causing a client device to display an indication of the respective prequalification result.
2. The method of claim 1 , further comprising determining that the applicant is suitable for borrowing from a lender of the plurality of lenders based on additional decision-making information and at least one of applicant preferences or filters applied by the applicant, wherein remaining lenders of the plurality of lenders are restricted from access to the additional decision-making information.
3. The method of claim 2 , further comprising authorizing, based at least in part on the respective prequalification result for each lender of the plurality of lenders, the applicant for a transaction associated with the product.
4. The method of claim 1 , wherein the applicant information further comprises an external record of a market source that the applicant traversed before reaching the server.
5. The method of claim 1 , further comprising selecting the plurality of lenders from additional lenders to pre-qualify another applicant by:
applying a lender pre-screening rule set based on at least one of Boolean logic operators or machine-learning logic; and
updating the lender pre-screening rule set by updating at least one of the Boolean logic operators, applicant information thresholds, or the machine-learning logic based on the respective prequalification result for each lender of the plurality of lenders.
6. The method of claim 1 , further comprising selecting the plurality of lenders from additional lenders to pre-qualify another applicant by:
applying the applicant information to a lender-specific neural network machine-learning classifier with backpropagation and at least one hidden layer, wherein user-provided information forms an input layer of the lender-specific neural network machine-learning classifier, and an output layer of the lender-specific neural network machine learning classifier contains two nodes, and wherein the another applicant is deemed to be suitable for borrowing from a particular lender of the plurality of lenders based on values of the two nodes.
7. The method of claim 6 , further comprising updating the lender pre-screening rule set by adjusting weights between the two nodes or additional nodes of the lender-specific neural network machine-learning classifier based on the respective prequalification result for each lender of the plurality of lenders.
8. A system comprising:
a memory; and
a processor coupled to the memory, wherein the processor is configured to perform operations comprising:
encrypting, by one or more computing devices, a plurality of rule sets for prequalifying an applicant to be stored within a lender repository of a jailed environment on a server, wherein a respective rule set of the plurality of rule sets is output from each trained machine-learning model of a plurality of trained machine-learning models, wherein each trained machine-learning model of the plurality of trained machine-learning models is trained to generate rules for a respective lender of a plurality of lenders, wherein, based on the jailed environment, the respective rule set output by each trained machine-learning model of the plurality of trained machine-learning models is accessible to only the respective lender;
receiving a request to generate a prequalification result for a product based on applicant information;
for each lender of the plurality of lenders and responsive to the request to generate the prequalification result for the product based on the applicant information:
decrypting, within the jailed environment and based on a respective lender-specific key, the respective rule set,
generating, in parallel within the jailed environment, based on the respective rule set and the applicant information, a respective prequalification result, and
causing a client device to display an indication of the respective prequalification result.
9. The system of claim 8 , the operations further comprising determining that the applicant is suitable for borrowing from the lender of the plurality of lenders based on additional decision-making information and at least one of applicant preferences or filters applied by the applicant, wherein remaining lenders of the plurality of lenders are restricted from access to the additional decision-making information.
10. The system of claim 9 , the operations further comprising authorizing, based at least in part on the respective prequalification result for each lender of the plurality of lenders, the applicant for a transaction associated with the product.
11. The system of claim 8 , wherein the applicant information further comprises an external record of a market source that the applicant traversed before reaching the server.
12. The system of claim 8 , the operations further comprising selecting the plurality of lenders from additional lenders to pre-qualify another applicant by:
applying a lender pre-screening rule set based on at least one of Boolean logic operators or machine-learning logic; and
updating the lender pre-screening rule set by updating at least one of the Boolean logic operators, applicant information thresholds, or the machine-learning logic based on the respective prequalification result for each lender of the plurality of lenders.
13. The system of claim 8 , the operations further comprising selecting the plurality of lenders from additional lenders to pre-qualify another applicant by:
applying applicant information to a lender-specific neural network machine-learning classifier with backpropagation and at least one hidden layer, wherein user-provided information forms an input layer of the lender-specific neural network machine-learning classifier, and an output layer of the lender-specific neural network machine-learning classifier contains two nodes, and wherein the another applicant is deemed to be suitable for borrowing from a particular lender of the plurality of lenders based on values of the two nodes.
14. The system of claim 13 , the operations further comprising updating the lender pre-screening rule set by adjusting weights between the two nodes or additional nodes of the lender-specific neural network machine-learning classifier based on the respective prequalification result for each lender of the plurality of lenders.
15. A non-transitory computer-readable medium storing instructions that, when executed by one or more processors of a device, cause the one or more processors to perform operations comprising:
encrypting, a plurality of rule sets for prequalifying an applicant to be stored within a lender repository of a jailed environment on a server, wherein a respective rule set of the plurality of rule sets is output from each trained machine-learning model of a plurality of trained machine-learning models, wherein each trained machine-learning model of the plurality of trained machine-learning models is trained to generate rules for a respective lender of a plurality of lenders, wherein, based on the jailed environment, the respective rule set output by each trained machine-learning model of the plurality of trained machine-learning models is accessible to only the respective lender;
receiving a request to generate a prequalification result for a product based on applicant information;
for each lender of the plurality of lenders and responsive to the request to generate the prequalification result for the product based on the applicant information:
decrypting, within the jailed environment and based on a respective lender-specific key, the respective rule set,
generating, in parallel within the jailed environment, based on the respective rule set and the applicant information, a respective prequalification result, and
causing a client device to display an indication of the respective prequalification result.
16. The non-transitory computer-readable medium of claim 15 , wherein the operations further comprise determining that the applicant is suitable for borrowing from a lender of the plurality of lenders based on additional decision-making information and at least one of applicant preferences or filters applied by the applicant, wherein remaining lenders of the plurality of lenders are restricted from access to the additional decision-making information.
17. The non-transitory computer-readable medium of claim 16 , wherein the operations further comprise authorizing, based at least in part on the respective prequalification result for each lender of the plurality of lenders, the applicant for a transaction associated with the product.
18. The non-transitory computer-readable medium of claim 15 , wherein the applicant information further comprises an external record of a market source that the applicant traversed before reaching the server.
19. The non-transitory computer-readable medium of claim 15 , wherein the operations further comprise selecting the plurality of lenders from additional lenders to pre-qualify another applicant by:
applying a lender pre-screening rule set based on Boolean logic operators or machine-learning logic; and
updating the lender pre-screening rule set by updating at least one of the Boolean logic operators, applicant information thresholds, or the machine-learning logic based on the respective prequalification result for each lender of the plurality of lenders.
20. The non-transitory computer-readable medium of claim 15 , wherein the operations further comprise selecting the plurality of lenders from the additional lenders to pre-qualify another applicant by:
Applying the applicant information to a lender-specific neural network machine-learning classifier with backpropagation and at least one hidden layer, wherein user-provided information forms an input layer of the lender-specific neural network machine-learning classifier, and an output layer of the lender-specific neural network machine-learning classifier contains two nodes, and wherein the another applicant is deemed to be suitable for borrowing from a particular lender of the plurality of lenders based on values of the two nodes.