James’s modeling module was built in close collaboration with risk professionals from all over the world. Its interface was designed under SEMMA’s framework to make sure the model building process is innovative without being disruptive.
Additionally, James aims to make sure risk professionals can get the most out of the data they work with. James incorporates state of the art machine learning techniques that allow you to optimize your model’s performance while ensuring its robustness. These techniques are also time-saving since they replicate hundreds of processes that usually have to be manually constructed.
James also allows you to try different classifiers - like the traditional scorecards/logistic regression or non-parametric models - without requiring a single line of code.
Once the model is created, James provides you tailored validation reports. In a few seconds, you can access a document that displays the model’s discriminatory power, calibration, and stability.
Every time a model is created, James prepares a document with all the most relevant statistics, so the user does not need to waste time calculating them for every experiment ran in our platform. These reports follow the directives given by the Basel Committee on Banking Supervision in order to make sure that they cover the needs of highly regulated lenders.
With the goal of facilitating the relationship between risk and IT professionals, we’ve built a deployment module. This module gives you the ability to deploy the model on Amazon Web Services or displays all the necessary information to set it up on premise.
However, we are aware that every lender has its own infrastructure. If none of our standard deployment solutions fits your requirements, we can easily generate custom made environments.
Current monitoring techniques are mostly reactive. We intend to change that by giving lenders real-time model auditing. On top of that, we’ve developed a system of alerts that will immediately inform you if your model’s performance is deviating from what was expected of it.
With these features, not only will risk professionals be allowed to proactively monitor the models they’ve built, they will also be able to rest assured that James will let them know if an unexpected event comes up. Having the monitoring and modelling modules in a similar interface gives the opportunity to quickly react and update models when required.
Does James provide alternative data?
No. James works with whatever data you provide. James is specialized in pattern detection. It does not add any data on its own.
Are your models blackboxes?
No. Scorecards and logistic regressions built in James are as explainable as those built in any other solution. It is true that models built using non-parametric classifiers are harder to explain. However, we still provide the stats required to validate them under Basel’s guidelines. On top of that, James has an add-on - we call it ‘Whitebox’ - that explains every decision made by these models.
Can I update the models myself or do I need your team to do that?
Our goal is to empower risk professionals. We continuously work with our clients to give them the autonomy to use James by themselves.
What happens if we don’t want to keep working with you?
Nothing. You lose access to James. If you wish to permanently end the relationship, we’ll delete any account and data linked to you. Yet, you maintain intellectual property over every model and report built during the duration of the contract.
I want to launch a new lending operation, but I have no historical data. Can you still help?
Yes. We’ve helped several lenders jumpstart their credit operations. We’ll set up a structure that will allow you to leverage machine learning as soon as you collect enough data to create your own models.
Do you sell scores or scorecards?
No. We can however work as an outsourced risk team. We are aware that risk professionals are in high demand and that building a team from scratch is no easy task. To help lenders tackle that, we offer our expertise to fill in the gaps, giving you the opportunity to focus on your strengths.