LEV SELECTOR
New York City, USA
Tel: 212-795-3979
Email: selector (AT) pipeline.com
web site: http://www.selectorweb.com
SUMMARY:
* Hands-on Architect / Database Architect / System Integrator / Project Leader / Programmer.
* Extensive financial industry background
(including data warehousing, database architecture, ETL, data migration and cleaning, Internet/database
applications, trading systems, server-side processing).
* Ph.D. - mathematical modeling and
computer simulations.
-----------------------------------------------
COMPUTER SKILLS:
LANGUAGES: Perl (also C/C++, Java (Sun Certified), JavaScript, Visual Basic, Pascal, Fortran)
DATABASES: Sybase, DB2,
Oracle, Microsoft SQL Server, MySQL, Berkley DB, MS Access
PLATFORMS: Unix (Solaris, Linux,
BSD, Mac OS), MS Windows
TECHNOLOGIES:
* Perl (database
maintenance, reporting, feeds, administration, CGI)
* DBArtisan, SQL Navigator, Erwin, Informatica (ETL)
* Apache/mod_perl
* HTML/DHTML,
CSS, JavaScript, Dreamweaver, Adobe PhotoShop
* Secure Crt Telnet, CuteFTP,
SVN, CVS, Perforce source control, GNU Make
* Java (Sun Certified), JDBC,
Java Servlets, JSP, WebLogic,
Eclipse
* XML/XSL, IBM MQSeries
* ASP, InterDev, and Front Page,
IIS
OTHER SKILLS:
* Advanced analytical skills (Ph.D. in Mathematical Modeling).
* Excellent Technical Architect and Database Architect. Analysis of customer environments, design of strategies and technical architectures. Projects' implementations. Excellent in documenting projects.
* Project Leader and Lead Tech. Excellent communication skills.
* Many years of teaching experience (computers, programming, sales/marketing).
* Real life sales and business experience.
* Financial industry experience.
-----------------------------------------------
PROFESSIONAL EXPERIENCE:
December 2007 - June - HSBC, http://www.assetmanagement.hsbc.com/ , Consultant.
Technologies: Perl, Sybase 12.5, Oracle 9, Windows Server, Control-M, Subversion.
Designed and implemented a framework for multiple feeds and reports for Asset Management group. Worked with business analysts and traders. Have successfully created the framework, wrote ~100 scripts and deployed/scheduled ~40 jobs. Some of these jobs are being used to move billions of dollars over periods of just few months. Some specific projects: Trade Feeds, Credit Risk Reports, Compliance & Audit Reports, Data Warehouse Loaders.
The framework is programmed using ActiveState Perl and many standard modules: DBI, DBD::Sybase, DBD::Oracle, Net::FTP, Net::SFTP::Foreign, Net::SMTP, MIME::Lite, Date::Calc, Archive::Zip, etc. The framework have utilities to do basic database work (including processing output of stored procedures with multiple result sets), ETL work (Extract,Transform, Load) including bulk loads/extracts, filewatcher (on local network or via FTP), file parsing (CSV split and join, delimited files, Bloomberg trade files and reports), working with dates, calculating business days using 2 or more different holiday calendars simultaneously in the same script, sending emails with attached data, logging, email alerts, scripts for recursive search and maintenance, generation of MS Office documents (Excel spreadsheets, MS Word documents).
I have configured all environments (PROD, DR, TEST) to use the framework. Wrote maintenance scripts to mirror important directories from PROD to DR using robocopy utility. Documented architecture, deployment and DR procedures, as well as specific scripts. Tested DR failover. Wrote multiple maintenance scripts (for removing old logs and data files, making backup copies of SVN repository, monitoring, etc.).
-----------------------------------------------
September 2007 - October 2007 - JPMorgan Chase, www.chase.com , Consultant.
Technologies: Sybase 12.5.3, Sybase Replication, Perl & Shell Scripts, XML Config Files.
As part of a consulting team from Hewlett Packard participated in a short-term project reviewing architecture of ETL processes (Extract-Transform-Load) and data flows in several datastores (ODS, PBDS, PCSDS) in Private Banking. The goal of the review was to find how the system should be changed to be capable to process more data (x5 and more) in the same time frame.
The system consist of numerous databases (one of the main Sybase servers has 114 databases, size of tempdb is 12.5 GBytes). The DBs are mostly loaded at night time using batch processes. Total load size is close to 100 million rows. There are tens of different sources, hundreds of feeds. Data mostly come in the form of flat files in different formats. Various technologies are used to generate and process these files: Mainframe Cobol extracts, Informatica, Ab Initio, Syncsort, Perl scripts, shell scripts, bcp, Sybase bulk API, homegrown java applications, XML, Sybase stored procedures, Sybase subscription based replication, Sybase Warm Standby replication, SRDF (Symmetrix Remote Data Facility) for data files replication.
Review has uncovered that current architecture doesn't scale as needed. Especially in situations when some of the feeds arrive late, thus causing many processes to run simultaneously and bringing the DB servers down. Specific recommendations were made to the management on how to rearchitecture the system. Here are some of recommendations:
* Remove replication where possible - use dual-load instead.
* Move calculations out of the database into the ETL layer.
* Convert complex chain of data flows and multiple ETL layers into a singe ETL layer.
* Upgrade to Sybase ver. 15.0.2 to take advantage of semantic partitionning to speed-up loads and processing.
* Introduce run-Ids and snapshots to make ETL jobs re-runnable and re-startable.
-----------------------------------------------
August 2006 - August 2007 - Merrill Lynch, www.ml.com , Consultant.
Technologies: Object-oriented Perl, Oracle 9, 10, Sybase 12.5, 15, NAS filers, Acopia switches, NFS, Unix (Solaris, Linux).
Global Banking and Investment Banking Group Technology - Technology Architecture Solutions and Services. In charge of Operations for Enterprise File System (EFS). EFS is a global file-system namespace designed to make development and deployment processes easy. It is maintained via Perl application with Oracle database as a backend. EFS was released into production in January 2007, and is growing fast as more developers from different business areas are coming onboard.
Projects:
* Compiling and integrating packages and libraries into EFS (for example, Sybase Open Client v.12.5 and v.15 for Linux and Solaris, miscellaneous Perl modules and unix utilities).
* Developing new code, testing deployment.
* Creating development utilities.
* Testing/tuning system performance on a Gigabit network. Writing necessary test scripts and conducting multiple tests in all 5 regions (US,UK,JP,HK,SG).
* Training new developers in the group and in new client groups.
* Working with Unix SAs and Storage teams in all regions to proper configure and maintain the system.
* Working with Monitoring Group on setting up proper monitoring procedures.
* Performing production releases.
* Maintaining documentation and making reports (for example, created a web site using Apache / mod_perl and HTML::Mason, populated another web site using Confluence Enterprise Wiki software).
-----------------------------------------------
July 2005 – July 2006 - JPMorgan Chase, www.chase.com , Consultant.
Technologies: Sybase 12.5, Perl 5.8, Solaris, Java (JSP).
Projects:
* Perl / Database work for JDS (Joint Distribution System). Database Architecture. Architectured and developed an ETL (Extract - Transform - Load) perl framework. Wrote multiple perl scripts to FTP, transform/clean the data, bcp in/out, doing custom loads and updates, generating audit tables and triggers, performing database maintenance and log maintenance. The framework included perl wrappers for proper error capture and reporting.
* Wrote many triggers and stored procedures. Some of the challenges included achieving good performance for Client Search screen and for multiple reports. The solutions included using Dynamic SQL, using string-suffix tables, using "update-with-variables" method.
* Wrote intelligent monitoring scripts, which periodically test and report the health of the application - and restart the application if necessary. These scripts were necessary because the Tomcat server is known for having periodic crashes, and these scripts have really decreased (virtually eliminated) their negative impact.
* Wrote scripts for data distribution and regression testing. These scripts are "working horses" of the project development cycle.
* Created necessary documentation for database architecture, and for Perl modules and scripts.
-----------------------------------------------
May 2005 - Alexandra Investment Management, www.alexandra.net, Consultant.
Technologies: MS SQL Server 2000, Perl, Erwin.
Project:
* Security Master database Architecture. Development of an ETL (Extract - Transform - Load) perl framework for S.M., documenting it, running test loads. The framework uses DBI ( DBD::ODBC ) to communicate with the MS SQL databases. It also uses BCP and FTP to move data. The system traps all errors, writes logs, and sends status emails and error alerts. Special effort was put into making the framework very simple to use and easy to integrate into existing enterprise systems.
March 2005 - May 2005 - Morgan Stanley, www.morganstanley.com, Global Prime Brokerage, Consultant.
Technologies: Unix (Solaris, Linux), Sybase, Perl.
Projects:
*
Development of a data-synchronization utility capable of handling big tables/loads (Sybase).
* Documenting the Entitlements system (jobs and interfaces) - Sybase.
-----------------------------------------------
February 2004 - March 2005 - CSFB,
www.csfb.com, Prime Services, New York, Programmer.
Development and support of an Operational Data Warehouse.
Technologies:
- Unix (Solaris), Windows
(NT, 2000, & XP).
- Sybase (v.11.9.2, 12.5) - Transact
SQL, stored procedures, triggers, import/export, real time feeds.
- DBArtisan, Informatica
(ETL)
- Perl, Bash, Shell scripts, CGI,
Java servlets (Weblogic), CVS version control, Control-M scheduler.
- HTML/DHTML (Dreamweaver MX), Javascript, Graphics
(Adobe Photoshop).
Some Projects:
* Development/maintenance of
the Prime Services Data Warehouse (PSDW) - a set of 4 databases on one Sybase
server. Main tables contain millions of rows (some - tens of millions). Hundreds
of tables. Three environments. Challenges include performance tuning, avoiding
blocking while running many concurrent processes accessing the same main tables.
Close to 200 jobs running during the day importing, exporting, processing data
(products, prices, positions, accounts, activity, rates, etc.). Most of the
jobs are working as ETL utilities (Extract, Transform, Load) written in Perl
or Bash. They use BCP, FTP, Sybase stored procedures, custom SQL, external file-based
cash. Some utilities are implemented in Informatica.
* Development of a big searchable
web site to keep documentation for the project. Netscape web server, Dreamweaver
MX, Adobe Photoshop, Perl for search engine script and for maintenance scripts.
* Tree diagrams showing
dependencies between the jobs. Custom Perl scripts parse the reports generated
from Control-M scheduler, map dependencies - and generate html pages. These
diagrams have proved to be very useful for support and design purposes.
-----------------------------------------------
December 2000 - May 2003 - Goldman Sachs,
www.gs.com, New York, Consultant.
Multiple financial applications and utilities: server-side,
GUI, feeds/integration, reporting, documentation.
Technologies:
- Unix (Solaris, Linux), Windows
(NT & XP).
- Sybase (v.11.5.1, 11.9.2, 12.0),
DB2 (v.6, 7.2) - Transact SQL, stored procedures, triggers,
import/export, real time feeds.
- DBArtisan, Informatica
(ETL)
- Perl, CGI, Java servlets (Weblogic),
Jython, Shell scripts, Visual Basic, CVS version control.
- HTML/DHTML, Javascript, Graphics (Adobe Photoshop).
- Siebel (v.6.3 and 7.5).
Some Projects:
* Owning and maintaining the
company's OE (Organizational Entities) database, and applications based on this
database. The database contains vital information (clients, accounts, contacts,
interests, employees, coverage, etc. - and their relationships). The work started
with cleaning the system of hundreds of perl and shell scripts, and hundreds
of stored procedures, documenting the project (creating a rather big local web
site with documentation). Multiple changes were made as required (both back-end
and front-end).
* Cleaning the data and developing
the tools for users to clean the data and monitor the data quality.
* Designing new scripts for
reporting and maintenance (including recursive "tree-reports").
* Analysis of the ways to move
the application into Siebel CRM system. Schema migration. Tools migration. Database
synchronization during the migration process. XML data exchange.
* Porting parts of the system
to DB2 (rewriting SQL from Sybase stored procedures to SQL for DB2 v.6 which
doesn't have T-SQL).
* Participating in writing some
java servlets for a new branch of the project (using WebLogic appl. server).
* Siebel: multiple screens,
bus. services and virtual components, feeds, documentation, maintenance.
* Siebel reports: weekly progress
report, daily budget report, daily data-quality reports.
* Reporting using Business Objects
and Data Warehouse database (DB2), big weekly report, contact interest queries,
performance tuning.
* Data quality monitoring. Multiple
reports and scripts to identify and correct the problems.
* Preparing data for Data Warehousing.
Using Informatica and custom perl-based feeds to do massive
updates (ETL - data Extraction, Transformation, and Loading).
* Live-update system for sub-account
linking. Sybase notification / perl.
* Sub-account linking and reporting
for Currencies and Commodities: design, creating feeds, multiple tables, stored
procedures, perl scripts, browser-based and Siebel-based tools.
* RepId maintenance tool (Perl,
Sybase/DB2, Siebel).
* Semi-manual OE-Merge process
(Perl, Sybase/DB2, Siebel, generating Excel-based reports).
* Complete redesign of the sub-account
linking process and tools to include history and purged sub-accounts. Performance
tuning.
-----------------------------------------------
July - October 2000 - digiTRADE - A Thomson
Financial Business Unit (www.digitrade.com)
New
York, NY. Consultant.
Financial application with Browser front-end,
Java Servlets and EJB middleware (Netscape Application Server) and SQL
database. Unix (Solaris). Technologies include: Java,
Java Servlets (Netscape Application Server 4), XML, XSL, XSLT, EJB (Enterprise
Java Beans), Netscape Enterprise Server, Perl, Korn Shell, SQL, Unify
database,
Netegrity site management,
HTML, DHTML, Javascript, Graphics (Adobe Photoshop).
*
Wells Fargo online
brokerage web site and U.S. Bancorp Investments, Inc. Online Investing web
site - helping on final stages of putting a project into production. Correcting
existing code (Java servlets, Javascript, HTML and graphics).
*
Working with client
on formulating the requirements for a new project ("Single Signon"). Development
of a authorization servlet prototype (parsing an XML token, getting a confirmation
from Netegrity SiteMinder).
*
Assisting and participating
in conducting a training devoted to using XML/XSL ( XSLT ) in Enterprise.
-----------------------------------------------
April - July 2000 - Morgan Stanley Dean
Witter, (www.msdw.com)
New
York, NY. Consultant.
Browser-based application with Java Servlets
(in Web Sphere), Netscape web server and Sybase database.
Unix (Solaris) and Windows NT platforms. Technologies include: Java,
Java Servlets (ServletExec, IBM WebSphere, VisualAge, Apache JServ), JDBC,
Perl, Javascript, HTML, DHTML, XML, XSL, SQL, Sybase.
*
Working with clients
on formulating the requirements for a new Portfolio Accounting system (eMSPA).
Designed architecture for the application. Implemented a prototype
of a browser front-end (multi-frame HTML with hidden frames, Javascript,
CSS, Per/CGI).
*
Working on shifting
the project from Perl/CGI to Java servlets. Writing Java servlets
for database access (JDBC), connection pooling, data caching, connection
with MQSeries, coding messages in XML. Testing servlets with different
Web servers (Apache, Netscape Enterprise) and Servlet engines (JServ, ServletExec,
WebSphere). Testing servlets in secure Client
Link environment.
*
Development of DHTML
templates and JSP pages.
*
Working on integrating
of a Scheme Editor into the eMSPA project.
*
Documenting the procedures
and libraries (local web site + javadoc)
-----------------------------------------------
June 1999 - April 2000 - Cantor Fitzgerald,
(www.cantor.com,
www.espeed.com)
New
York, NY. Programmer.
Maintenance and redesign of a large distributed
trading systems. Unix (Sun OS) and Windows NT platforms.
Technology includes Apache with mod_perl, Cgi / Apache::Registry, Perl
DBI, Sybase, SQL, HTML, DHTML, JavaScript, graphics, Java. Some projects
are listed below:
*
A fix for a history
system allowing brokers to search and edit instruments in the Sybase database
via browser interface.
*
Design and implementation
of a software documentation web site (in 6 months - 150 MB of compiled
and structured documentation and schematics, navigation).
*
Architectural design
of a history spread system: a server listening to a trading
system, processing data and storing it in a database, and
a client
allowing brokers to search the database and pull info about particular
issues and history reports (bids/offers, etc. + calculated spreads + graphical
charts).
*
DTS (Distributed
Trading System) maintenance
*
Implementing and
documenting of administrative procedures for DTS (Distributed Trading System)
servers setup and troubleshooting.
*
Design of an administration
and distribution system for DTS (Distributed Trading System). Functionality
includes distribution of files among multiple servers, remote installation/backups,
remote restart of the servers as well as urgent rollback/restart utilities.
Technologies include browser-based interface with password protection and
data validation (HTML, JavaScript, Perl/CGI, Apache/mod_perl, suid scripts).
*
Implementing and instructing
on using a Perforce source control system
(used from Unix command prompt as well as from Windows client).
*
Analysis of bottlenecks
in existing DTS (Distributed Trading System) and formulating the ways to
dramatically improve its performance and reliability.
-----------------------------------------------
1998 - May 1999 - Waterhouse Securities,
Inc., (www.waterhouse.com)
New
York, NY. Internet Programmer.
Design of
enterprise size scalable web
sites with load balancing, business applications and security.
Windows NT and Unix (Solaris, BSDI) platforms. Technology includes
web servers (Apache, Stronghold, Netscape Enterprise 3.6, IIS-4), Netscape
Application Server, CGI scripts (Perl, C), HTML, JavaScript, VBScript,
ASP, graphics, Oracle, SQL. Some projects are listed below:
*
Testing system load.
Upgrading the hardware for better performance and scalability.
*
Architectural redesign
of the site ( waterhouse.com )
(including routing, firewalls, load balancing, web servers, back end processing).
*
Perl spider-scripts
were developed to automate the process of cleaning the site and correcting
bad links.
*
Customization of
webBroker application for Jack White & Co. Conversion of approximately
600 HTML pages and templates. Special spider-scripts were written in Perl
to automate the process. Also numerous graphics were customized.
*
Automating of data
transfer between the web site and Oracle database (Perl scripts, cron job).
*
Design of interactive
forms-applications using Perl, HTML templates, cookies. Reverse IP
verification for proper processing of client's information.
*
Design of secure
logon with encrypted cookie.
*
Design of a new version
of ad-tracking software (including server part and interface for marketing
dept.).
*
Adapting an interactive
financial calculator (an object-oriented-style Perl/web software) to work
on the main corporate site.
*
Design of methods
for effective and safe testing and deploying of new software on the production
site.
-----------------------------------------------
1998 - Pierce Technology Corporation,
(www.pierce.com) New York, NY. Technical
Consultant.
*
An
interactive
corporate web site for Pierce Technology Corporation was developed.
Technology included HTML, graphics, Perl, CGI scripts and DBM database
on Unix server.
*
July 1998
- a short Internet publishing project with Time Inc. New
Media. Development of templates for a Story Server. Technologies
included Story Server, Tcl/Tk, Perl, Sybase database, Dynamic HTML.
-----------------------------------------------
1994-1998 - Infolink International, Inc.,
New York, NY. Project Leader.
*
Web design and promotion
for business clients. Various
interactive web applications were developed to represent businesses on
the Internet, capture prospects, automate and stimulate sales process.
Work included the analysis of customer environments and developing strategies
for applying internet technologies. Work included HTML editing (with
tables, frames, forms, graphics optimization and animation), Perl CGI scripting,
working with database files (Unix DBMS), database user interface, main
search engines registration, banner exchange. Technology used:
HTML, Perl-5, Unix DBM files, IE-4.0, Netscape Communicator-4, Eudora Pro,
Crt Telnet, Cute FTP, Adobe PhotoShop, Ulead GIF Animator, search engines
submission tools.
* Complete database design and implementation for Infolink International, Inc. The database system was developed to automate order taking and accounting. The database was working on a local network and was used to process orders for numerous service products for up to 3,000 clients ordering through 300 agents. Work included the design of the database's structure, user interface, screens, menus, reports, documentation, fax distribution system, and accounting procedures. Main modules - Orders and Agents. System included numerous forms for order entry and adjustments, search function, flexible data import/export, calculating commissions using different compensation schedules and manual adjustments. System was printing reports for agents. Technology used: MS Access, Access Basic.
-----------------------------------------------
1991-1994 Columbia University, (www.columbia.edu , Prof. L.Liebovitch), New York. Staff Associate.
* Mathematical modeling of dynamics of organic molecules. To speed-up the calculations, programs were written in C and distributed to several unix computers in different labs. Operations were controlled from one desktop Macintosh computer. Results were fed from unix to Mac and custom programs (Mac C++ and IGOR software ) were used to automatically graphically process and print-out the data.
-----------------------------------------------
1981-1991 National Cardiology Research
Center, (www.telemed.ru/cardio)
Moscow, Russia.
* Project originator and lead developer - real time data acquisition and processing in neuro-physiological experiments.
* Project originator and lead developer for hardware and software design of a portable myograph-neurograph for clinical use. (Big project. First version of software was ~25,000 lines of code, software technology used - Pascal, Assembler, C).
* Mathematical simulation of nerve impulse generation and propagation along C-fibers. Partial differential equations, Hodgkin-Huxley model, Crank-Nicolson method, modified Runge–Kutta method.
* Teaching programming and electronics to graduate students.
-----------------------------------------------
EDUCATION:
1985-1988 Graduate School of Moscow Phys.-Tech.
Institute Ph.D. in Biophysics (computer simulation of nervous
coding).
1975-1981 Moscow Physics & Technology Institute.(
www.mipt.ru
)
MS in Electronics and Automation, Majors in computers,
electronics and biophysics, Diploma - computer simulation of nerve activity.
-----------------------------------------------
REFERENCES:
Available upon request.