| Since June 2023: |
Coremont, London; Head of the Equities and Commodities Quants
|
| March 2018 - June 2023: |
HSBC, London. Front office Exotic Equity Derivatives Quant - Director. I have been assisting the Equity Derivatives trading by building trading tools (mostly in Python) and performing analysis in order to better understand, anticipate and manage risk; while conducting the trading activity. For some projects i have been acting as the primary contact in the quant team, to interact with other divisions (including Product Control), and facilitate the usage, within the bank, of the quantitative tools we create. As the leader for many projects I have been guiding, training, and monitoring junior staffs in order to deliver number of key projects, for the Equity Derivatives Business. Some of the projects i lead include:
- As the leader of a team of three colleagues i have organised the design, implementation and delivery of a Python tool to mark the correlations within the Equity Derivatives business. The delivery of this project positively impacted the business by saving us millions of pounds per year
- I implemented a framework to estimate the Implied Volatility Bid-Offer PnL Reserves (implied volatility cost of unwinding a portfolio or a position, in a stressed market). I did a presentation to all the trading desks, worldwide. It was a success. For this framework I designed and proposed an innovative model to model the Black and Scholes implied volatility Bid-ask Spreads
- I designed and built a tool used by the HSBC top management to lively monitor the Portfolios Inventory (notional, gnbv, stochastic volatility reserve, vega, epsilon, cega), for the Equity Derivatives business, worldwide
- I designed and built a Neural Network and xgboost based object to generate signals in order to implement systematic trading strategies, by predicting the implied volatility move, for a given maturity horizon. The tool will be used by (Equity) QIS (Quantitative Investment Strategies); and some trading desks. I fully implemented the tool in Python, from scratch. I used Keras and Tensorflow to set up the Neural Network part of the tool. I can easily reproduce it
- I supervised an intern to implement a generic outlier detector in Python. The tool is now used at HSBC within the implied volatility update process. The tool is fast (less than 30 seconds to classify a volatility surface), generic and works with both multi-dimensional and mono-dimensional data. I can easily reproduce it
- I wrote a theory about the replication of the ith principal component using exactly i input variables. The paper has not been published. I back tested (in Python) for the implied volatility surfaces analysis, taking each point (moneyness, maturity) as an input variable. The results were very satisfactory. I was able to replicate the first principal component as a function of only one volatility point (moneyness, maturity); the second principal component as a function of exactly two volatility points; and the third principal component as a function of exactly three volatility points. In back test the correlations between the principal components and their respective replicators were all higher than 97%. I named this theory the Minimum Principal Component Analysis (MPCA). It can be seen as a neat improvement of the Sparse PCA. One useful application of this theory is that the trader can track the main drivers of his PnL, and can also replicate those to hedge or anticipate his PnL future moves. Working Paper;
- I designed and built (from scratch) the HSBC Python framework to compute risk metrics (mainly Gap Risk) displayed to the trader to manage their margin loans book
- I wrote the theory and built a framework to perform the Profit and Loss (PnL) explain related to any scenario of the risk factors driving the book. The framework uses some sophisticated formula, and is fully built in Python. Data are loaded from a Mongo data base, KDB data base and some HSBC XML-structure data storage. The tool is used all over the bank (London, Paris, New York and Hong Kong) to assess the toxicity of each derivative product in any trading book; and therefore gives directions for the pricing policy and the risk management. The tool is accessible through a Flask (web) interface that i built with the assistance of my intern
- Business As Usual: daily maintenance and enhancement of the C++ pricing library, plus assisting the business (trading) to mathematically understand and find quick solutions to some urgent problems to be solved in order to make new deals or monitor existing deals
|
| May 2012 - March 2018: |
Bank Of England, London; In charge of most of the quantitative developments within the PRA and for the Bank of England Markets activities:
- Modeling, pricing and hedging simulations on products from different asset classes: Variable Annuities with local volatility; Swaps; Swaptions; Autocallable bonds
- I did a review of the pricing theory following the 2007-2008 crisis: Multi-curves Interest Rate construction, including Cross Currency Swap; Pricing of a collateralised
contract with option of choosing between two collaterals to post; calculation of the unilateral and the bilateral CVA. Working Paper;
- I designed and built the tool used by the Prudential Regulation Authority (PRA) to generate the adverse scenario sent by PRA to all the banks they rugulate, for the trading book stress testing;
- Ahead of the "Brexit" referendum i designed and built a model for assessing how big the sterling would depreciate against other major currencies, if the britains were to vote for 'leave'.
The model predicted well. Working Paper;
- I designed and built a tool for interpolating and extrapolating the bond's yield curve (term structure). I used the two factors vasicek model for the short interest rate. Working Note
- I implemented the free boundary SABR model to provide the Bank of England Monetary Policy Committee with the Swap Rate probability density function implied from the Swaptions prices quoting in the market. Working Paper
- I produced a framework explaining how to implement the Markowitz Modern Portfolio Theory for any investment portfolio size, and for any types of equality and/or inequality constraints. Working Paper;
- I privately (out of my employment context) built a generic tool in VBA, to perform the Principal Components Analysis (PCA) on any time series of multi-dimensional variable. Available here
|
| July 2010 - May 2012: |
Quant within the Valuation team at Societe Generale, Paris; In charge of the Collateral
- Implementation of a pricer for some credit derivatives: CDS, Bond and CDO. Working Paper;
- Implementation of a Gas and Electricity pricer;
- Implementation of a pricer for options on forward bonds and options on Constant Maturity Treasury (CMT). Working Paper
|
| April 2008 - July 2010: |
Quant within the front office equity derivative research team at HSBC,
Paris:
2 factors and 1 factor stochastic volatility models, simple Heston:
- Calculation of volatilty over premium on some exotic products;
- Implementation of Local-Stochastic Volatility (LV-SV) model Working Paper;
- Calculation of the term structures of: volatility of volatility; spot-variance correlation; asymptotic skew. Internship Report
- Calibration.
|
| September 2007 - July 2008: |
Assistant of teaching at Ecole Polytechnique,
Palaiseau (France): Giving extra teaching in applied Mathematics, to first and second year engineer students
|
|
|
| April - September 2007: |
Research internship on the topic: Adaptation of the local volatility calculated from a parametric implied volatility, for a better replication of Variance Swap prices, at BNP Paribas, Paris
|
