Portfolio


This page shows a selection of past client engagements. Most of our work on validating models / algorithms qualifies as company confidential. This means that in the below cases no client names are mentioned.
However, the highlighted cases should provide sufficient insight in the type of engagements, approach and results achieved.
Of course, we look forward to discuss how we can help you.


Large International Bank


Issue

A large international bank needed to demonstrate on short notice that their model for loans to Large Corporate clients was still fit for purpose.

Approach

A dedicated Aisure team backtested the model under time constraints during the year end holidays

Results

A validation report demonstrating the continued sound model performance, ready to submit to the bank supervisor- with minimum interference on the bank staff/resources.

What can we do for you?

Large domestic bank


Issue

The trading desk of a large domestic bank was short on staff to do the required annual review of their trading risk model.

Approach

An Aisure team did the annual review using historical trading data and model documentation delivered by the bank. Additionally, the aisure team interviewed key users on the functioning of the model.

Results

Model review finished on time demonstrating the trading risk model (and trading desk) was sufficiently in control.

What can we do for you?

Bank Insurance Company


Issue

A Bank Insurance company wished to apply for an internal model approach for their regulatory capital. However, the technical documentation for their most material model- mortgages- was not up to regulatory requirements.

Approach

A joint aisure/client team reviewed and improved the technical model documentation, including clear references/ evidence to the regulations.

Results

Model documentation that was compliant with regulations and ready for submission to the supervisory authorities.

What can we do for you?

Large bank Finance department


Issue

A model to value/price illiquid loans was enhanced using machine learning techniques. However,it was not clear why the performance improved.

Approach

A dedicated Aisure team independently validated the Machine Learning model, looking specifically at data, methodology, robustness and explainability.

Results

A validation report that gave confidence in the enhanced model for deployment in quarterly reporting/disclosures to investors.

What can we do for you?

Domestic insurance company


Issue

A domestic insurance company wished to build an internal validation department as part of an application for Solvency2.

Approach

Aisure provided training and hands-on support for this newly established department.

Results

A jump start for a new department by quickly building experience and becoming an internally trusted business partner.

What can we do for you?