XVA Research & Development
HSBC
Senior AI Ambassador (HSBC)

Youssouf
H. Kerzika

XVA Quantitative Researcher
contact@kerzika.com

Seven years at HSBC, Exiom, and Mazars in quantitative finance covering rates, FX, equities, and inflation across investment banking, model design, and production engineering. Built XVA calibration frameworks, simulation engines, and monitoring infrastructure to production grade: from mathematics through to automated pipelines and AI diagnostics.

What I build
YK
Youssouf Kerzika
What I Build
Primary Expertise
01

XVA & Counterparty Risk

Development of end-to-end cross-asset XVA frameworks on Rates, FX, Equities, Inflation and Credit. Extend the framework to quantify and aggregate counterparty risk through netting and collateral structures. Establish model assumptions, limitations, and diagnostic tests around these dimensions, continuously refining methodologies to address instability, bias, and quantifiable model shortcomings.

02

Derivatives Pricing & Calibration

Design and implementation of cross-asset stochastic models including Hull-White 3F for rates, Heston for equities, correlated FX dynamics, and local volatility surfaces, with end-to-end calibration pipelines from market data to production-ready parameters.

03

AI-Augmented Risk Infrastructure

LLM-based diagnostics identifying calibration instability and parameter drift, BiGAN-driven synthetic scenario generation to enhance stress test coverage, and automated monitoring pipelines with assumption-level alerts.

96%
calibration parameter estimation runtime reduction
HSBC
95%+
reduction in model monitoring cycle time
HSBC
30%+
extension in stress test coverage via BiGAN synthetic scenarios
Exiom Partners
Stack
Python C++ Git GCP CI/CD Jenkins BiGAN LLM Pipelines QuantLib PyTorch
Regulatory
SA-CVA BA-CVA FRTB IMM
Methodology

I documented the core derivation, calibration logic, and production architecture in the writing page.

Full writing & research xvafoundations on GitHub
Writing & Research

A technical introduction
to XVA and derivatives.

Two series: Interest Rate Modeling builds the curve and calibration foundations, then XVA Computation takes those building blocks into exposure simulation, aggregation, and counterparty risk valuation adjustments.

Most of the institutional knowledge in XVA lives in internal presentations that never see daylight. This material deserves a permanent, public record; every formula derived, every chart reproducible, every code snippet tested.

Browse all writing
Interest Rate Modeling
Interest Rate Modeling

From curve bootstrapping and swaption volatility surfaces through stochastic short-rate models, Hull-White calibration, and Monte Carlo pricing of interest rate derivatives.

4 chapters live
XVA Computation
XVA Computation

Exposure simulation, netting-set aggregation, CVA computation, and beyond: taking the rate models into production-grade counterparty risk valuation adjustments.

3 chapters live
Work With Me
Open to the right conversation.

Available for consulting on XVA methodology, model calibration, and AI-augmented risk infrastructure. Also open to speaking engagements and desk education at banks, trading houses, and quantitative funds.

Get in touch LinkedIn Profile