Data Support

* ******* ****, ********, **, ***** Phone: 203-***-****, E mail: ********@*****.***

Ilya V. Feldshteyn

Summary ● 20+ years of data analysis experience in different applied areas with 10+ years of experience in finance.

● Team player with good interpersonal skills and experience in small group coordination.

● Data analyst with vision and interest in new developments in the field. Professional

experience

2014 present: Booz Allen Hamilton, New York, NY, USA Senior Lead Scientist

Key responsibilities:

● Provide consulting in data science and direct client’s developers and IT groups.

● Work on integration of Hadoop based tools (Hortonworks and Pivotal environment) into client’s infrastructure.

● Work on internal research projects in the field of data science applications and algorithms developments.

● Provide mentoring for entry level data scientists. Achievements:

● Developed working prototype of portfolio risk analysis tools including data accumulation, portfolio adjustment, scenario approach, risk quantification.

● Developed Hadoop and Solr based framework for call center messages processing with the purpose of identifying sensitive topics. The approach is based on frequency analysis combined with artificial topic generation (Latent Dirichlet Allocation LDA). Prepared suggestions for future developments.

● Developed initial revision of entity matching library. The core element is machine learning module that builds matching function based on multiple fields includes neural networks and random forest approaches for different types of data.

● Worked on reviewing new possibilities provided by quantum computers in collaboration with D Wave Systems. Studied partially connected Hopfield network where connections correspond to D Wave qubits matrix structure. Confirmed that this type of connections works for basic recognition tasks.

● Participated as a technical mentor in joined project of BAH and the United States Holocaust Memorial Museum. The goal of the project is to use data science to run screening of available open sources to predict genocide. Provided support for data collection group.

● Worked with summer interns provided help in setting up machine learning environment, choosing analytic approaches, programming and running basic studies in text analytics and neural networks.

2013 2014: The Falconwood Corp, New York, NY, USA

Quantitative Analyst

Key responsibilities:

● Support portfolios of commodity futures.

● Build, back test, and support futures related indices.

● Execution system programming.

● Prepare research materials for executives reviews of statistical arbitrage and machine learning in applications to commodity futures trading. Achievements:

● Worked on predictive models for commodity futures. Predictive models include proprietary machine learning framework, and the statistical arbitrage approach developed in Alta Advisors and adapted for small universe of commodity futures.

● Implemented arbitrage strategy for trading sector ETFs vs some commodities.

● Reviewed analytic approaches to using Reuters market news sentiment feed and macro economic factors as inputs for predictive models and for portfolio risk monitoring. This type of data sources supposed to be external elements when building affinity matrix

the element important in risk analysis.

2004 2013: Alta Advisors LLC, New York, NY, USA

Alta Advisors worked as a sub adviser for Paloma Partners, Millennium Partners, The Falconwood Corp, The Horton Point.

Data Analyst

Key responsibilities:

● Support and improve statistical arbitrage platform.

● Collect, process and analyze SEC filings. Develop predictive model and support trading strategy based on it.

● Improve portfolio optimization approaches to address multiple risk related constraints and estimation of total cost of execution.

● Develop and support intraday tick based market simulator used for order execution back testing.

● Support market data accumulation and cross verification with various data sources, for different markets (equities and futures), different time scales and of different nature

(prices, corporate actions, identifiers, fundamental data).

● Work with vendors on integration of their products into Alta’s infrastructure.

● Develop and support execution system, trades/portfolio processing and accounting

(including multiple strategy environment).

Achievements:

● Statistical arbitrage platform worked on portfolio of US equities up to $350MM, containing up to 2000 positions with 3 4 days turnover rate. The predictive model used factorization with principal component analysis and non linear regression to build forecast. The biggest challenge was to put forecast into context of generation of multi factor neutral portfolio. The original universe of 3000 3500 stocks resulted in quadratic optimization problem with ~9,000 variables and 16,000 20,000 constraints.

● Worked on estimation of market impacts for different order execution tactics. Ran extensive analysis of intraday market move patterns associated with trading activity.

● Built data collection and processing environment that worked with multiple market data sources: Bloomberg, S&P CapitalIQ, RealTick, Reuters. Tested several other data vendors. Worked on customization of several third party products, including coordination of internal development and vendors solutions.

● Developed predictive model based on SEC filing data (10 K, 10 Q, 8 K). The model was based on automated balance sheet analysis and long term patterns retrieval. Analyzed price patterns between sequential filings. The strategy worked on a small portfolio in 2006 and first half of 2007.

● Intraday market simulator was built and worked on ~1.5TB of tick data. Calibrated execution model. The model provided estimations of average prices, fill rates, and execution time period. The model showed good correspondence to real trading results.

● Built and supported trading infrastructure that incorporated several independent components data modules, intraday real time feed connector, intraday decision module, and execution system based on Tethys Technology solution. Different parts of the infrastructure used MS SQL Server, proprietary data formats for quick access to market data, modules coded in MATLAB, C#, C/C++, Python, Transact SQL.

● Made improvements in portfolio optimization. Improvements addressed multiple issues in risk management. Reviewed multi horizon optimization showed that at some conditions multiple forecasts on different scales lead to quadratically constrained optimization problem.

● Worked on adjusting statistical arbitrage system originally developed for big universe of US equities to the small universe of commodity futures front contracts (portfolio of 15 20 positions).

Alta Advisors was a small company. Daily work included active collaboration with colleagues and providing technical and methodological consulting in my field of expertise. 2001 2004: Cimmetry Systems Inc., Montreal, Quebec, Canada Senior Software Developer

● Worked in the field of reverse engineering of CAD systems, low level design, development and implementation of parsing algorithms. 1999 2001: Japan Advanced Institute of Science and Technology, Tatsunokuchi, Ishikawa, Japan

Research Associate

● Led a team that developed and tested on line system of experimental data accumulation and analysis.

● Developed the method of optimization in some numerical computations in molecular dynamics simulation on parallel super computer (Cray T3E). 1995 1999: Keldysh Institute of Applied Mathematics, Moscow, Russia

Scientist

● Developed mathematical methods of data analysis and diagnostic software in medicine

(electrogastrography).

● Worked in the field of chaotic dynamics: theoretical analysis of multi dimensional data reconstruction methods, forecasts in time series, and diagnostics of complex systems. Developed applied methods of distributed systems analysis.

● Worked as a member of research team for Russian Ministry of Emergency Situations. Developed methods of technological catastrophe forecasts and statistical analysis. 1993 1995: Institute for Problems in Mechanics, Moscow, Russia Scientist

● Developed the methods of data analysis in mechanical engineering in application to express diagnostics of friction junctions. Nonlinear data analysis and system state estimation.

● Developed the methods and software for optimizations in anti vibration systems. The system developed is used for analysis of efficiency of wind protection devices for high voltage and optical cables lines.

Education 1998 PhD in applied mathematics, Institute of Applied Mathematics, Moscow, Russia 1993 MS in applied mathematics, Moscow Institute of Physics and Technology, Moscow, Russia

Technical

skills

Languages: Python, MATLAB, C/C++, Java, SQL

Technologies: OOP, databases, APIs of financial data providers. Hadoop and related packages (Hive, Solr, Sqoop)

Contact this candidate