Quantitative research, portfolio management

REGINALD BROWN

C 310-***-**** ********@*********.***

PROFESSIONAL EXPERIENCE

PALOMA PARTNERS, Greenwich, CT

Quantitative Analyst, 2016-2017

Invited to join a pre-existing group of three people. Implemented a market neutral medium frequency (few days) statistical arbitrage strategy for US equities as a compliment to their longer term strategies. Responsible for daily portfolio management, alpha modeling, and risk modeling. Unlike the instances described below, all data was provided by existing infrastructure. Management selected the stock universe, and an execution strategy developed by JP Morgan execution services.

Software tools: Matlab in a Linux environment.

The full alpha model has 11 distinct signal classes, each uses different data, or processes the data using a different methodology. Each class has three or more time scales from fast to slow.

The trading universe consisted of a monthly changing list of 500 names, selected on the basis of low volatility.

INTELLECTUS PARTNERS, San Francisco, CA

Quantitative Analyst, 2015-2016

Implemented, backfilled, and maintain a SQL database for quantitative analysis and modeling. Developed three strategies for managing client investments. The first strategy optimizes a portfolio of hedge funds on a quarterly basis. The portfolios are specific to the individual needs and risk tolerance of each client. It includes risk management, and transaction costs to control over trading positions. In testing, the strategy outperformed the current methodology by 50%. The second strategy is a long only equity strategy that generates a portfolio on a monthly basis. In testing, and after fees, it outperforms the S&P500 index by 70%, and outperforms the two external long only firms by 40% each. The third strategy allocates client funds to investment styles (bonds, long only, long/short, multi-strategy funds, etc.) according to target return, cash flow, and risk. The strategy then selects specific allocations for a portfolio of approximately 20-30 instruments.

Software tools: Matlab, SQL, VB.Net, Bloomberg API.

A proprietary GUI is used to maintain the database.

Proprietary GUI’s are used to run each of the three strategies.

ORBITAL INSIGHT, Palo Alto, CA

Contractor, 2015

Orbital Insight is a big data startup that interprets satellite images, and produces time series data that is marketed to financial firms. Demonstrated the efficacy of car count data by developing an alpha model, with two predictive signals. Implemented a strategy that uses the alpha model to trade quarterly report announcements for 20 retail firms. During testing, the strategy traded 323 out of 455 announcements with a hit rate of 56%, and a t-statistic of 2.5. Orbital Insight and this strategy are discussed in articles published in Bloomberg news, and bank white papers.

Software tools: Matlab

QUANT GLOBAL CAPITAL ADVISORS, New York, NY

Portfolio Manager, 2013-2015

Implemented a refined version of the US equity market neutral strategy developed at Telemetry Capital Management. Responsible for all portfolio management, software development, alpha modeling, risk modeling, universe selection, and execution. Market data obtained from DTN, static data obtained from FactSet, analyst data obtained from Thomson-Reuters. Created a unified SQL database for all data storage, and proprietary GUI’s for data capture, trading, and backtesting. Created a proprietary execution management system using the RealTick API.

Software tools: Matlab, SQL, C++, VB.NET, MarketQA, FactSet API, DTN API, RealTick API.

The trading GUI allows the portfolio manager to control all positions and/or risk (when necessary).

The full alpha model has 15 distinct signal classes; each uses different data, or processes the data using a different methodology. Each class has three or more time scales from fast to slow.

The trading universe contains all members of the Russell 1000 index.

CONCEPT TAPESTRY, Santa Clara, CA

Contractor, 2012-2013

Provided strategy research, best practice advice, and technical expertise to Concept Tapestry, a start-up fund using machine learning to trade intraday on India’s BSE and NSE exchanges.

Key Accomplishments:

Performed basic research, and wrote reports for industry best practices and strategy evaluation.

Reports included:

oFeature extraction for alpha models and optimal forecast horizons for alpha models.

oBest practices for data infrastructure.

oDefining and controlling risk for portfolio and trade based strategies.

oExpected profitability/risk analysis for proposed strategies.

oPerformance evaluation statistics for existing strategies.

THINK TRADE/TELEMETRY CAPITAL MANAGEMENT, San Diego, CA

Portfolio Manager 2008-2011

Charged with the mandate of building a one-week holding period statistical arbitrage (U.S. equity market neutral) strategy to leverage outside investors.

Key Accomplishments:

Single handedly planned, performed basic research, developed and implemented in software all infrastructure, models, universe selection, and portfolio management for the strategy.

Infrastructure included a proprietary unified SQL database (all tables and stored procedures for data and model parameter insertion, storage, and extraction), a trading platform with GUIs to backtest and trade the strategy, market data capture via Bloomberg Server API, and static data loaders.

Models included a proprietary alpha model methodology, factor risk models, nonlinear transaction cost models, and a proprietary portfolio optimizer.

Strategy marketed as a separate self-directed LLC (Telemetry) with $3 million in seed capital. Maintained 10% return on capital and Sharpe ratio of 1.5 in paper and live trading.

Software tools: Matlab, SQL, VB.NET, C++ and Bloomberg API.

TELESIS CAPITAL, Los Angeles, CA

Director of Research 2006-2007

Served as a quantitative risk specialist. Performed basic research and implemented risk and performance analysis factor models to control risk and style exposure in the portfolio of hedge funds. The techniques included symbolic dynamics for the factors and regression to fit the models.

Software tools: Matlab and SQL.

BARCLAYS CAPITAL, London, England

Associate Director 2004-2006

Implemented a Pan European statistical arbitrage strategy, while serving as the sole quantitative analyst on the Delta-1 trading desk.

Key Accomplishments:

Created a unique medium frequency (2-week equity market neutral) strategy with face value of $40 million USD, a return of 4-5% on notional value, and a Sharpe ratio above 2.

Wrote all infrastructure software (except market data capture) for alpha modeling, risk modeling, loading static data, portfolio optimization, and portfolio management.

Selected and maintained the trading universe.

Recognized for establishing the only profitable portfolio based quantitative strategy on the desk.

Software tools: Matlab, VB.NET, Bloomberg API and VBA.

CREDIT SUISSE FIRST BOSTON, London, England

Vice President 2003-2004

Performed for all quantitative research and development as a member of a 3-person team. Conceived and wrote all software for a new intraday statistical arbitrage strategy for Pan-European equities.

Key Accomplishments:

The software included alpha models, risk models, and portfolio optimization for the strategy. These efforts resulted in a fivefold increase in net profit, and a threefold increase in Sharpe ratio, over the initial design for the strategy.

My efforts increased net profit by 400% and Sharpe ratio by 200%.

Recruited by Barclays as a portfolio manager for a new three-continent Delta-1 effort.

Software tools: Matlab.

TELIC MANAGEMENT, Greenwich, CT

Senior Quantitative Analyst 2002-2003

Acquired, evolved, and integrated new trading techniques that supported and maximized modeling efforts of a small firm within the Paloma Partners group of funds.

Key Accomplishments:

Created a mean reversion signal based on no-arbitrage principles as a single class of predictive signals for multifactor alpha models.

Developed a momentum signal obtained from quarterly financial statements as a second class of predictive signals.

Recruited by Credit Suisse to serve as lead quantitative analyst for a new Pan-European statistical arbitrage strategy based in London.

Software tools: R, Perl, SQL, and C.

PREDICTION COMPANY, Sante Fe, NM

Senior Research Analyst 1999-2003

Performed basic research, and wrote software, for the U.S. medium frequency, and long term asset management teams.

Key Accomplishments:

Delivered a proprietary portfolio optimization algorithm that increased live trading net profit by 100%.

Delivered an alpha-modeling algorithm that reduced model complexity by 85% while maintaining identical performance. The alpha modeling approach was later introduced and adopted as the precedent algorithm by Prediction Company.

Created the first set of four alpha model factors from S&P Compustat data for a new long term asset management strategy.

Software tools: S+ and C.

COMPUTATIONAL PROFICIENCIES

Programming Languages: MATLAB, R/S+, VBA, VB.Net, C#, Java, SQL, C/C++, Python, Perl, Ruby, and Maple. Multi-threading in C/C++, VB.Net and C#.

Software Packages & Data Vendors: NAG, Visual Studio, SQL Server Management Studio, ADO.Net, Access, Market QA, MQA Pervasive, FactSet API, Bloomberg Server API, RealTick API, Thomson Reuters IBES, COMPUSTAT, TAQ.

Operating Systems: Windows, UNIX/Linux

EDUCATION

College of William and Mary, Faculty Physics Department and Faculty Department of Applied Science

Visiting Scholar MIT Boston, NRL Washington DC, Institute Applied Physics Nizhny Novgorod, DTU Copenhagen

University of California, San Diego, CA, Senior Researcher

University of California, Berkeley, CA, Hass School Financial Investment Technology Program

University of Maryland, College Park, MD, Ph.D., Physics/Dynamical Systems (#1 ranked US program)

University of Puget Sound, Tacoma, WA, BS Mathematics and BS Physics

ACADEMIC AND PROFESSIONAL ACHIEVEMENTS

Series 65 license (expired)

CAREER Award, National Science Foundation

GOALI Grant, National Science Foundation and Hewlett Packard, Principal Investigator

Air Force Office of Scientific Research Grant, Principal Investigator

Ralph D. Myers Award, University of Maryland

Recognized for 38 published papers (15+ dedicated to forecasting and parameter estimation).

Contact this candidate