Computational Machine Learning Engineer



The XVA / Scarce Resources team is part of the Global Market Division (GMD). This 30 people strong team entails 3 sub-teams:
• 1 trading team: based in Paris, London and Hong Kong in charge of pricing XVA and hedging to reduce PnL volatility.
• 1 Quants team based in London and Paris.
• 1 XVA Strategy Projects and Transformation team (XVA ST) based in London and Paris.
In the framework of major regulatory changes, the mandate of the team is to:
• Reinforce Bank risk management
• Help reach and maintain the right balance between Meeting accounting & regulatory constraints whilst remaining competitive
• Optimise scarce resources like Risk-Weighted Assets (RWA), Leverage Ratio…
• Manage defaults
The mandate of the quant team: is to produce quantitative modelling and innovative solutions for XVA, Counterpart Risk, Collateral
and Credit topics. The quant team regularly interacts with a broad scope of internal clients:
• XVA and Scarce Resources desk for XVA pricing and modelling
• Risk department for Internal & Regulatory CCR, Accounting XVA, and SIMM
• Collateral desk for discounting, SIMM and IMVA with CCPs
• Trading and Risk Management for Credit derivatives.
The quant team closely works with the business to study and assess the models’ behaviour and performance. It also plays a significant
role in several strategic XVA and RWA projects by producing computational blocks using cutting-edge modelling and implementation
techniques to ensure the bank can cope with the increasing list of regulatory measures (XVAVaR, SACCR, FRTB-CVA …) and metrics
needed to manage our XVA reserves properly. As such, the quant team will be strongly involved in the Smart XVA Project.
The quant team continuously builds and upgrades XVA libraries and platforms to implement regulatory changes in an optimised
architecture. The team is also actively participating in developing the Collateral management platform for CCP and EMIR Initial Margin
and working on various FO and Risk systems migration projects.
Smart XVA Project: in June 2021, “Direction Generale” has validated and sponsored the so-called “Smart XVA and Scarce Resource
Optimiser” project, which consist of building a neural network using AAD sensitivities and leveraging some machine learning techniques
to compute XVA and possibly other metrics later on once the neural network is made available as a service.
The immediate use cases are to:
• Be able to compute XVA fast enough so that it is compatible with E-trading.
• Be in a position to generate on-demand via a “Scarce Resource Optimiser tool” to be built, portfolio of trades that are optimal
from a scarce resources perspective to proactively offer those; packages mutually beneficial to our counterparts.
More use cases are expected once the 2 first ones are addressed (FRTB, accounting XVA computation etc.)
This project is closely intertwined with the XVA Quant library development, and for that reason, it requires knowledge primarily on the
quant side and then data science / machine learning.


Key Responsibilities for the Computational Machine Learning Engineer
• Thanks to close interaction with other team members, high Financial Modelling and C++ programming skills.
o Quickly master XVA implementation in the XVACCR Library.
o Assimilate the AAD methods recently implemented to compute XVA sensitivities to initial Market Data.
• Propose and discuss various solutions using Neural Networks to speed up XVA computation drastically
• Implement and produce sensitivities specific to the Smart XVA project.
• Use a C++ API of a proven Machine Learning Library to Implement a Neural Network:
o Stable (Training always providing a solution) and with a good asymptotic behaviour
o Apply AAD to differentiate the Neural Network.
o Train the Neural Network using the sensitivities produced for the project.
o Isolate the issues that need to be resolved, to make programmes more effective.
o All ML steps and components are implemented in a dedicated dll.
• Use trained Machine Learning to compute XVA.
• Propose and discuss various solutions to use Neural Networks in E-trading and RWA optimisation.
• Demonstrate end-to-end understanding of applications (including, but not limited to, the ML algorithms) being created.
• Interaction with IT to industrialise various lots of Smart XVA project.
• Efficiency and accuracy of developments
• Reactivity in the function of supporting users
• Innovation in models and numerical techniques
Legal and Regulatory Responsibilities
• Comply with all applicable legal, regulatory and internal Compliance requirements, including, but not limited to, the London
Compliance manual and the Financial Crime Policy.
• Maintain appropriate knowledge to ensure to be fully qualified to undertake the role. Complete all mandatory training as required
to attain and maintain competence.
Key Contacts:
• Other members of the XVA quant team
• XVA Trading and Strategy & Transformation Team
• IT
• Innovation Teams
• Risk Teams


Poste : Computational Machine Learning Engineer

    * Champs obligatoires

    Votre CV doit être au format PDF

    Revenir aux offres