Ilyes Kerkeni

FX desk strat

Building trading strategies using Machine learning and reinforcement learning

Gap year internship Global market division (GMD), USD vol desk -Collecting and analyzing data from US insurers, forward rates, implied volatilities, and index values to anticipate and explain insurers’ trading activities. -Building and maintaining new models.

-Worked with both machine learning models and multivariate rough volatility models, gaining expertise in the randomized signatures (rSig) pricing model and the quadratic Gaussian (qGauss) models. -Developed rich models that remain numerically tractable for the fast generation of autocallable prices and sensitivities. -Conducted calibration of the models to var swaps, covar swaps, vol swaps, and options, ensuring that the models were accurate and effective in generating these sensitivities.

A two-step semi-parametric estimation approach for dynamic conditional value-at-risk (CoVaR) which is used to define reserves for a given financial entity to limit loss overruns when a system is in distress. key words: Value-at-Risk VAR, CoVaR, Bootstraping, GARCH models.