|
Mike's Resume
Michael P. Wilson
Skillset Overview:
High-level OOA&D. Implementation in C++. Design Patterns and
Pattern Languages. Specialties in: high-performance library
development and integration, legacy code remediation, programmatic DB
access (embedded SQL, etc) including database persistence.
Environment, Language and Platform highlights:
C++ Perl SQL Linux XML
Win 32 ODBC Java DB2 UDB
11/2000 - 5/2002 TradeWeb LLC
- Implemented protocol engine for the Financial Information Exchange
protocol (FIX version 4.2)
- Designed and implemented a Message-Oriented-Middleware system from
the ground up featuring hot service discovery and adaptive routing,
and pluggable client-side protocol adapters (including FIX 4.2, FIX
4.3, FIXML, and a proprietary XML-based protocol.)
- build full heavily templated baseline toolkit library including
threading support, logging facility, adaptable streaming classes,
pluggable content filters, factories, shared resource managers, ODBC
wrappers, etc.
(All in pure ANSI C++ with minimal OS specific code wrapped tightly in
toolkit libraries mentioned above.)
11/98 - 7/2000 IBM
Consultant (Programmer/Architect)
Designed and developed middle tier of the massively distributed
Anti-Virus Immune System. Implemented within the scope of a web
server, the Gateway Array was implemented to handle the load of a
full-scale virus epidemic. Speed, flexibility, stability, and
scalability (platform independence) were all critical.
The Gateway was written completely in ANSI C++, using ODBC to access a
DB2 database. Among the more interesting features are: support for
SSL, database connection pooling, massive logging and tracing,
intelligent caching of resources, and hot-reconfiguration of gateway
machines.
Most of the code written for the Gateway array was abstract enough to
be used in the implementation of many other components in the Immune
System as a whole.
7/97 - 10/98 Aetna/US Healthcare
Consultant (Programmer/Architect)
Architected (and led development on) a rewrite of Quest, a 400-user
application for the automated fulfillment of questionnaires
accompanying RFPs, requiring many independent subsystems (written
wholly in ANSI C++) including:
- An RTF Parsing engine with callback hooks for complete customization
of state transitions, data sources, etc. (Later recompiled into a DB2
stored procedure on AIX.)
- Text Analysis Engine that parses a questionnaire document and yields
the "outline hierarchy" of questions as individual fragments (in plain
text and RTF using the above library.)
- An advanced Data Conversion application that cleansed and normalized
'1 record per file' data into 15 independent database tables (or the
equivalent file structure.) Including translation of single-character
lookup codes to exploded text values with db-friendly primary keys.
(App also runs maintenance 'diffs' for weekly updates from legacy
system.)
- Statistical Analysis Engine and reporting facility to provide
management with sales-performance feedback of fulfilled RFPs (which
Questions and/or Answers tend to generate closed contracts, etc.)
Designed and implemented Document Management Library (for use in
multiple applications) in C++ to handle persistent storage and
retrieval of Documents in ODBC accessible databases (using ODBC
library below), file systems, or other storage mediums in a way that
was completely transparent to the application developer. The core
capabilities of DocMgr allow for processing and composition of
composite documents, network-optimized caching, and support for pre
and post process plug-ins (e.g. encryption, compression engine,
document analysis and indexing etc.)
Assisted in release of "Field Proposal System" during Quest downtime
by integrating ODBC Library as a means of abstracting the data storage
layer, removing more than thirty C++ classes from the existing
application.
1/97 - 7/97 Concept Information Systems
Programmer / Analyst, Architect
Developed a Data Warehouse Synchronization system which cloned logical
access-specific subsets (such as for a particular sales region) of a
corporate data warehouse on to local user databases. An NT server
application performed the extract, and a client-side component
recreated the database schema and imported the data provided onto the
client PC. The Synchronizer is completely independent of the RDBMS
implemented on the client or server side.
Encapsulated the ODBC 2.0 API in order to implement the Synchronizer
project. All extract and some insertion methods were encapsulated as
required. The library supported dynamic SQL selects, returning
on-the-fly type specific bound data. DDL for recreation of selected
columns was generated as an automatic feature.
Architect & Senior Developer of an OLAP decision support system in
Visual Basic and C++, using effective Design Pattern techniques
despite VB5's inherently weak object model, making the project a
manageable mid-range (75k lines) Object-Oriented product. Guided the
project team through OOAD principles and techniques.
4/95 - 11/96 Walter Karl Inc.
Programmer / Analyst
Developed a client/server data repository and research application,
using Rational Rose and SqlWindows, Delphi 1.0, and Visual C++ for the
front end, and the SqlBase database engine on the server side. This
required a custom-built ad-hoc SQL query engine (Delphi). A C++
application parsed the data from EDI into a normalized relational
format for importing into back-end database. A text search engine
was developed for ad-hoc keyword, phrase, and word proximity
searches.
Built a client/server data-entry and lookup database application for
tracking of client data tapes with an automated export and upload of
data to a mainframe tape tracking system, kicking off an import
procedure online.
Performed numerous feasibility studies to determine corporate IT
re-engineering strategy. Provided continuous feedback through
extensive oral and written reports to management throughout this
period.
4/94 - 4/95 List Maintenance Corporation
Programmer
Responsible for converting client data to and from customized formats.
Developed a suite of tools to automate this procedure, increasing
throughput and accuracy by over %75.
Designed and implemented an application suite to automate internal
conversion and analysis processing for a single client. Simple
heuristic AI techniques were used to eliminate data redundancy and
optimize postal costs. Custom files were automatically created for
multiple destinations (TRW, Equifax, mainframe upload, WANG upload).
Original "by-hand" processing took two to three weeks of R&D time.
Resulting process took less than one-half hour of processing by an
account executive, significantly reducing processing costs and
removing developers from the loop entirely.
Education
9/87 - 6/88 Clarkson University Potsdam, NY
9/84 - 6/87 The Marvelwood School Cornwall, CT
|
© Copyright 2003 Michael Wilson.
Last update: 10/26/03; 19:36:58.
|
|
