Sign in

Quantitative Portfolio Manager/Researcher

Greater London, NW6, United Kingdom
March 13, 2018

Contact this candidate

Mohamed El Kourdi, PhD

Flat **, The Luminaire, **3 Kilburn High Road, NW6 7FA, London, UK

Tel: +44 (0) 73 42 80 08 42



To obtain a position with a well-established company that provides a highly intellectual and collaborative work environment, along with the promotion of a continuous learning paradigm.

Skills/Achievements Related To The Job Spec

NB: The following models have been conceptualized using state of the art advanced statistical, econometric, and information theoretic machine learning techniques. The development has been carried out through employing different programming languages (Python, C#, R, Java). Most of the models/techniques have been developed from scratch (not reliant on specific package) and adopted in hedge funds for systematic quantitative trading/risk management.

Systematic Quantitative Strategies/Alpha Signals/Pricing (Global Managed Futures) (Horizons: 1, 5, 10, 20, 65 days):

Regime Switching Directional Strategy Based on a Smooth Kurtosis Proxy: A Dynamic Correction of Momentum Factors Risk Exposures. (Bayesian Regression) (Traded: +2 Sharpe)

Traditional Trend Following (Autoregressive Bayesian Neural Network)

Global Macro Exploiting Inter-Market Relationships and quantitative proxies for inflation, economic growth, and interest rates (Reinforcement Learning)

Latent Factor Models (PCA and Clustering)

Value strategies (Stochastic Reinforcement Learning and Neural Networks)

Defensive Strategies (Sampling Theory and Pseudo Polynomial Time Dynamic Programming)

Smart Betas (Ridge, Lasso, and Elastic Nets)

IR Swap Term Structure: A Non-Linear Bayesian Approach

Portfolio Allocators/Optimization:

A Dynamic Programming Approach to Multi-Period Geometric Mean Portfolio Optimization: Beyond Mean Variance.(Levenberg Marquardt/Quadratic Optimization on Belman Equations)

A Genetic Algorithm for Stochastic Portfolio Optimization through Epi-Graph Reformulation of non-linear constraints.

Deploying Black Litterman for Multi Assets Managed Futures allocation.

An Efficient Method for Tackling the Curse of Dimensionally in Co-Kurtosis(Tensors) based Portfolio Allocation.

Volatility Estimates:

A Tail Risk Robust Two Factor Volatility Model Using Kalman Filter and the Statistical Stylized Facts of Range and Half-Normal-Distribution.

A Non Parametric Stochastic Volatility Model using Cornish Fisher Transform (Kernel Regression)

A Novel Statistical Proxy for a Smooth and Efficient Kurtosis Risk Estimate.

Empirical Modelling of the Non linear Relationship between Kurtosis and Volatility.


Bayesian/James Stein Correlation Shrinkage Based on Fisher Space Transform.

Correlation Filtering: Information Theoretic Hierarchical Agglomerative Clustering.

Power Mapping Versus Random Matrix Theory.

Market Microstructure/Slippage:

A Novel Market Impact Cost Model (Incorporating a Power Law Decaying Factor of Lots as a Function of Tick Volatility Levels).

A Market Impact Driven Execution Trading Strategy.

High Frequency Strategies (ticks, 1s, 1min, 5min, 10min, 1hr).(Genetic and RL Algorithms)

Volatility and liquidity relationship dynamic using a market maker’s risk aversion model

Stakeholders Management:

Frequent interaction with institutional investors and investment committees through presenting research findings in a higher level of abstraction.


July 2017- 31/12/2017

Company: MET Traders, London, UK

Position: Quantitative Portfolio Manager

Programming languages: Python, C#.

Managing and trading a proprietary systematic managed future (CTA) Portfolio (Commodities, Fixed Income, Stock Indices, FX) and setting up the quantitative arm of MET traders.

Global macro and Latent Factor trading strategies using advanced statistical and machine learning techniques, along with state of the art volatility forecasting models, correlation filtering, geometric return portfolio optimization

Interest Rate swaps term structure: A Bayesian approach based on Nelson Siegel predictors.

Conceptualizing five new systematic Fixed income strategies (Carry, Value, Quality, Global Macro, Directional)

March 2016- July 2017

Company: Liquid Capital Markets, London, UK

Position: Senior Quantitative Analyst/Portfolio Manager

Programming languages: Python, C#, R.

Fully responsible for setting up quantitative/systematic managed futures (CTA) trading strategies from ideas generation/design to implementation/production.

Portfolio management, design and implementation of a novel systematic medium term frequency strategy (Sharpe =2.1) for trading managed futures across all the asset classes (19% live track record from Dec 2016 to Mai 2017). Advanced statistical inference and risk management techniques have been used including an enhanced Bayesian regression technique, two factor volatility forecast, higher order moments estimators, correlation filtering, and geometric return portfolio allocator. Python 2.7 has been used for mathematical modelling, visualization, and production code (all the mathematical/machine learning models are developed from scratch rather than using package libraries).

Design, implementation of an adaptive latent factor model trading strategy based on PCA component scores of the sample correlation matrix and Akaiki information criterion. Simulation and back-testing have been carried out in diverse range of portfolios (60 futures contracts). Python 2.7 and C# have been used as programming languages

Design, implementation of a data driven inter-market relationships trading strategy using a developed Bayesian Neural Network. Asset classes include currencies, fixed income, commodities, and stock indices. Python 2.7 has been used as a programming language.

Working on consolidating latent the factor model strategy through using an agglomerative clustering model in order to capture the inherent non-linear risk factors that govern the impulse mechanics of assets returns. Python 2.7 has been used as a programming language.

Setting the direction of potential research opportunities in systematic medium term strategies, along with the presentation of research findings to the management board.

Design and development of a systematic high frequency trading strategy using reinforcement learning. Python 2.7 has been used as a programming language.

November 2008- July 2014 (1 year non-compete expired 21/07/2015)

Company: Altis Partners Ltd [Jersey, UK] (Global CTA)

Position: Quantitative Researcher

Altis Partners is a systematic CTA hedge fund (1.6 Billion AUM at peak) which trade a diverse range of future markets (Interest Rates, Bonds, commodities, Equities, FX) based on sophisticated forecasting and robust risk management techniques.

Programming languages: C#, R, Python, Matlab, SQL, ASP.Net

Designed and developed a novel systematic quantitative trading strategies and risk management models using advanced mathematical and Machine Learning techniques.

Ideas Generation and leadership of data science and quantitative trading projects.

Developed methods for making inference and discovering the hidden patterns from large data sets.

Designed and developed a forecasting framework (Bayesian Regression, Bayesian Neural networks, Reinforcement learning, evolutionary learning, kernel regression, Lasso, PCA) for predicting financial markets over different time horizons (from scratch using C#)

Designed and developed a novel short term trading strategy through candlesticks state encoding scheme patterns generated from stochastic reinforcement learning and genetic algorithms, along with risk adjusted returns forecasts inferred by learning on states/actions values space using Bayesian regression.

Designed and developed medium and long term trend following strategies using new indicators and Bayesian Regression

Formulated a new market impact cost model based on the limit order book and market microstructure dynamics, along with trade execution strategy.

Formulated and implemented a novel stochastic volatility forecasting model by incorporating higher order moments risks, along with two factor volatility model that is robust to short term shocks

Designed and developed latent factor models for forecasting risk adjusted returns based on the principal components analysis (PCA) scores of sample correlation matrices and Akaiki Information criterion.

Solved highly complex quantitative trading problems by exploiting theories from mathematics, physics, information theory, machine learning, and econometrics. such projects include Trading horizons coherence using Ising model, grey theory for forecasting volatility, and many other projects

Designed and developed quantitative inter-market relationships trading strategies across all the asset classes (FX, Equity, Commodities, Bonds, Short interest rates)

Designed and developed novel sample correlation filtering methods based on Information theoretic clustering mechanism, random matrix theory, and power mapping.

Designed and developed the risk management platform which consists of dynamic and multi-period portfolio allocators, risk attribution, VaR, linear and non-linear risk factors (Clustering), Portfolio optimizers.

Consistently improved the existing trading strategies through incorporating new trading indicators, improved volatility and correlation forecasts, new forecasting paradigms

Presentation of research findings to institutional investors, along with portfolio management and trade support

February 2008- October 2008

Company: Global Investment Research Ltd [London, UK]

Position: Quantitative Analyst/Developer

Global Investment Research Ltd was a start-up hedge fund trading future and forward contracts through all the asset classes

Programming languages: Java, C#, Matlab, SQL Server

Designed and implemented Cascade correlation neural network and autoregressive neural networks for forecasting risk adjusted returns

designed and implemented a portfolio optimizer based on Newton method and epi-graph reformulation

Implemented API for data feeds, database connectivity and the IT infrastructure needed by an early start-up hedge fund, along with the Implementation of risk management components including VaR, CVaR, CDaR


January 2006-March 2008: PhD in Computer Science (machine learning) from Staffordshire University: Full International Scholarship (tuition fees + living expenses), Stafford, UK.

January 2006-March 2006: Postgraduate Certificate in Research Methods (PgCRM) from Staffordshire University, Stafford, United Kingdom.

July 2003-August 2003: Summer School at the Technical University of München: Funded by the German Academic Exchange Service (DAAD), München, Germany.

July 2003: Master of Science in Computer Science (Artificial Intelligence: Machine Learning) from Al Akhawayn University (American based system): Funded by R&D Morocco, Coresoft SARL, and Al Akhawayn University, Ifrane, Morocco.

December 2000: Bachelor of Science in General Engineering (Major in software engineering and Mathematics) from Al Akhawayn University, Ifrane, Morocco.


Programming Languages: C#, Python, R, Java, Fortran, C, R, Matlab, TCL/TK, Assembly language, WML, XML, JavaScript, VBScript, XPath, XSLT, Prolog, Lisp, ASP.Net, JSP.Net, SSRS, WebMatrix, Source control(svn, git), unit testing.

DBMSs : Oracle8i, SQL Server, MS Access, MongoDB, and DB2

Software Architecture & Design: Distributed Multi-Tier, MVC Design, OO and function-oriented analysis and design.

Software Modeling: waterfall, Boehm (Spiral), and Rational Unified Process.

Networking : TCP/IP, IEEE 802.3/4/5, X.25, ISDN, Frame Relay, xDSL

Design & specification languages: UML, Z, ER model, and Natural.

Enterprise architecture tools: ARIS toolset, Casewise


Mathematical/ Machine Learning: Bayesian Regression/inference, Neural Networks (Bayesian, Recurrent, Deep), Kernel Regression, Reinforcement Learning, Dynamic programming, Decision Trees, Clustering, PCA, Numerical optimization (Levenberg Marquardt, Newton, gradient), Information Theory, Parametric and Non-Parametric Statistical inference, GARCH, Kalman Filters, Markov Bootstrapping, Association Rules, Fuzzy Logic, Text Mining, Information Retrieval, Decision Theory, Stochastic Calculus.

Systematic Quantitative Trading Strategies (All Asset Classes): short term, trend following, regime switching, volatility risk hedging strategies, adaptive Latent factor models, inter-market relationships and macro strategies.

Risk Modelling: non parametric stochastic volatility, multi-factor volatility models, correlation risk, correlation filtering, higher order moments, and money management strategies.

Market Microstructure: market impact modelling, optimal execution algorithms, high frequency trading.

Asset Allocation: Mean-Variance, multi-period geometric means maximisation, Black-Litterman and dynamic programming.

Portfolio Management: VaR/CVaR, risk factors, diversification metrics, and stress tests, scenario analysis.


M. El Kourdi, A. Bensaid, and T. Rachidi,"Automatic Arabic Document Categorization Based on the Naïve Bayes Algorithm", in Proc. of COLING 20th Workshop on Computational Approaches to Arabic Script-based Languages, Geneva, 23 August, 2004.

M. El Kourdi, T. Rachidi, A. Bensaid, A. Chekayri and M. Mhamdi. "A CONCATENATIVE STATISTICAL APPROACH TO NON-VOCALIZED ARABIC ROOT EXTRACTION”, in Proc. of the Sixth Conference on Language Engineering, Cairo 6th/7th December, 2006.

M. El-Kourdi, H. Shah, A. Atkins, “A Framework for knowledge discovery in EA, Journal of enterprise architecture (JEA), Number 4, Volume 3, November 2007.

H. Shah and M. El Kourdi, "Enterprise Architecture Frameworks ", IEEE IT professional Journal, Volume 9, pp36-41, 2007.

T. Rachidi, O. Iraqi, M. Bouzoubaa, A. Ben Al Khattab, M. El Kourdi, A. Zahi, and A. Bensaid, "Barq: distributed multilingual Internet search engine with focus on Arabic language," In proc of IEEE Conf. on Sys., Man and Cyber., Washington DC, 5th to 8th October, 2003.


15/11/2010-16/11/2010: Finance with R Programme, UNICOM Seminars and OptiRisk Systems, Brunel University, Uxbridge Campus, MiddleEsex UB9 4UDA, United Kingdom

19/5/2009: Introduction to commodity futures and options, NYSE Liffe in Conjunction with K2 London Ltd, Cannon Bridge house,, 1 Cousin Lane, EC4R 3XX, London, UK

2007: ARIS Toolset, IDS Sheer, Manchester, UK.

Apr 1999: Oracle Designer/Oracle Developer 2000, OmniData, Morocco.


English (Fluent)

Berber (Native)

Arabic (Fluent)

French (Fluent)

German (Basic)

Spanish (Basic)

Contact this candidate