The founder, Jolyon, having briefly considered taking a PhD after completing his postgraduate studies at the University of Cambridge, decided to transition to business and finance.

To gain general exposure to these areas, he joined EY and qualified as a Chartered Accountant whilst working in the audit department.

Jolyon spent the first couple of years auditing a wide range of different companies in varying industries. However, once his Chartered Accountancy exams were behind him he spent his spare time teaching himself financial mathematics and how to code in C++. He quickly became the ‘go-to’ person for derivative valuation and review work in the capital markets group.

He earned himself a prestigious posting to the risk and analytics group in the New York office. While there, he worked on complex credit structures and learned how they operated and how to model them.

Working on these complicated structures whetted his appetite for more, and Jolyon decided to move to the origination and structuring side of the business rather than being an advisor.

After five and a half years at EY, he left and joined a credit trading desk at Enron in London. This began a 14 year stint working first for Enron and then a couple of banks, followed by a hedge fund and finally raising capital with a colleague for their own structured credit business.

Because he was both a ‘quant’ and a Chartered Accountant – a fairly rare combination – Jolyon was able to move seamlessly between high level external and internal meetings and model development. He would give presentations to clients and comprehend highly technical business, legal, regulatory and accounting issues and immediately understand if there were any related model implications. As a result, technical issues didn’t ‘fall through the cracks’ between the business and the model development teams thus avoiding misunderstandings and delays.

His time in the city covered the financial crisis of 2008 / 2009 and, after that, the demand for the type of structures that Jolyon worked on reduced. Following the last structure he completed in 2015 / 2016 he was considering his next challenge.

He was aware of neural network models and was looking for an excuse to spend time working on them when an opportunity came his way in 2017 to assist an old colleague who was the co-founder of a data extraction company, Semantic Evolution. During his brief time there, Jolyon realised that this was an area that he felt he could meaningfully contribute to and, because of his finance background, he was drawn to the reasonably uncrowded space of deep learning for business data applications.

Up until this point, he had become extremely adept at building highly complex models but for a very specific purpose – that of credit derivative trading and structuring. He was excited about the prospect of becoming proficient at building a model that would appeal to almost any industry with almost limitless applications.

In 2018 he founded AIBMod and on a full-time basis threw himself into the pursuit of developing such a model with various objectives laid out here – something he has now achieved.

Jolyon and a colleague were offered the opportunity to start their own credit structuring business, Martello, with an initial investment of $55M. They built a small team and in 2010 managed to obtain the first post-financial crisis provisional credit rating for a Credit Derivative Product Company type entity.

Martello was chosen by a major European bank as the preferred solution to deconsolidate a bad bank from the balance sheet – the bank didn’t expect to find a solution that delivered the extensive benefits their structure offered.

In 2014 / 2015 Jolyon, on his own, conceptualised and structured a credit skew index trading entity liaising with a cross-border team of external lawyers and accountants to establish Earlsferry, a securitisation vehicle, that could issue listed securities and trade up to EUR600M gross notional of CDS vs index. This project had almost no quantitative element demonstrating his ability to manage highly technical (legal, regulatory, accounting, tax and associated cross-border issues), but not specifically quantitative, projects.

Jolyon was on the trading floor when the tragic events of September 11th 2001 occurred and it became clear that the valuations team could not simply revalue the book – instead, it was a complex operation that took many hours which frustrated the traders.

In his spare time, he began building in C++ a complete risk management framework with pricing, VAR, stress and scenario testing for the entire credit derivative book comprising single name CDS, First To Default and Single Tranche synthetic CDO tranches.

Aside from the valuation speed issues, this was hugely important to the desk as they had very simple gross VAR numbers which gave no benefit for hedging name exposures and it was beginning to become an issue.

Upon discussing the single tranche legal terms with the trading desk’s legal team it became clear that the tranches’ cash flows were not correctly modelled by the existing valuation tools.

In one and a half months the model was complete and it had two immediate consequences: the VAR number which was based on historical revaluation was greatly reduced giving more capacity to the trading desk and, secondly, it created $7M of positive trading P&L as the tranches were now correctly modeled.

Jolyon had also found computational efficiencies which greatly reduced the time required to calculate deltas and jump to default analysis meaning the trading desk had an incredibly useful and flexible ‘what if’ tool across the entire derivative book.

Years after he left Enron he found out that the receivers sold the source code to another London based financial institution for $2M.

Whilst in New York in 1999 with EY, they won a unique piece of work from the consortium of banks that bailed out LTCM. EY had to verify LTCM’s portfolio valuation a year after the bailout before a repayment. He was handpicked to verify the derivative part of the book as he was a fully qualified chartered accountant and derivative audit experience from London. He spent two weeks in Greenwich Connecticut reviewing the derivative portfolio and discussing his findings with the risk managers.

Whilst in New York in 1999 with EY, because of his experience building quick equity derivative Monte Carlo models in C++, he offered to help the team that was trying to build a capital model for a client attempting to launch the first Credit Derivative Product Company. The model was reviewed by the rating agencies as it was used to issue counterparty and debt ratings.

Despite having had no prior experience with credit derivatives or capital models, in one week he managed to create a working prototype in C++ that was considerably faster than anything produced before.

When he was due to return to the UK, EY recruited a financial mathematics postdoc researcher from a well known quantitative US University to take over the project. Six months later Jolyon was urgently called back to NY as his replacement was unable to develop the model at the same pace and the client was becoming frustrated with the lack of progress.