Company had over 1 million documents relating to FDA submission, company wanted to image and cross-reference documents.

Worked with third party and senior IT executives at company to strategize methodology for archiving documents. Designed system architecture, sourced hardware and software, supervised building and integrating system. Within three week of design, had system built and operational. Utilized network of high-speed scanners, Solaris and Novell servers, Optical storage, and ICR software, along with Informix database. Successfully archived and indexed over 1 million documents in three-month period.

Problem:

Company had DOS based system doing daily administration and NAV calculations for over 100 Mutual Funds. 7 Server Novell System used Btrieve Record Manager, had about 11,000 Btrieve files. System at capacity and number of processes not optimal. Large data loads took hours to complete.

Re-architect system to use IBM RS6000 Servers, IBM SSA Storage, and Informix database. Clients converted to Windows, critical batch processes moved to AIX to harness server power. System design and conversion took six months. Actual Data Conversion completed over a weekend, new system functioned perfectly first day, all processes completed ahead of schedule. New architecture allowed proposed system expansion, at peak close to 500 funds serviced. New server based data loads took minutes.

Above system was generating 10-15,000 reports per day, close to 50,000 reports at month or quarter end. Sophisticated document management system was needed. Number of major Document Management vendors submitted multi-million dollar proposals. When reviewed with Business users proposed systems did not fulfill needs to schedule, and index reports. Proposed systems required reports to be imaged and indexed after generation, workflow was ungainly, proposed systems could not handle projected volumes in timely fashion.

Designed report scheduling, optical storage, and retrieval system, processed reports in a batch mode as reports created by source system. All reporting moved to AIX, and batched nightly. Result: removed client network traffic, and freed up workstations (and daily printer battles). All reports immediately available for viewing when accountants arrived for work, saving hour(s) per day per accountant. All historical reports online, and readily accessible. Average seek times for daily reports - seconds, for history less than one minute. System implemented with trial group for one month, within three months rolled out to whole company. At peak system batched 15,000-20,000 reports per night on normal business day and 50,000 per night at month/quarter end. System implementation costs (including Optical Storage Hardware) less than $250,000.

System had a need to provide NAV and other data in wide variety of formats to feed multiple different client, and internal systems. Whenever change was required to NAV reports or new format was needed, programming resource was required, along with scheduling of production move. Led to unacceptable delays in implementation.

Provided a data driven report-writer style interface to allow Business Analysts to create new data export formats. Allowed for almost instantaneous turn around of system changes. Also standardized on data interchange tool, and tied tool to above functions allowing analysts to specify data interchange processes without programmer involvement.

Problem:

Company had major growth was using a Novell based Accounting System, system was highly customized for companies distribution business. Needed to build more robust system, to allow for additional growth and increase performance.

Lead role in designing new system using "off-shelf" GL accounting module integrated with custom developed AP Module, and mainframe AR module. Using clustered NT servers, and MSSQL database, built 2 tier Client/Server AP/GL system to support companies multi-billion $ business, harnessed power of SQL RDBMS to move nightly batch cycles from desktop to server.

Company wanted to Upgrade System to move nightly processing into an "n" tier model to harness performance of multiple servers.

Redesigned, and implemented, changes to system to use "n" tier model, built in multiple fail-over and redundancy. Resulting in major performance improvements.

EDI transactions were limited to 999 lines per check due to limitations of interim systems. Company wanted to have as many as 25,000 transactions per check.

Designed and Integrated new module into system to handle EDI transfer utilizing Java bridge between existing application modules, Database, and MQ Series messaging, allowing 9999 transactions per check (Bank Limitation).

Problem:

Company had older legacy system, needed to upgrade both Oracle Database, and major parts of code base. Many issues with current code base including, poor transaction models, inefficient locking, leading to single threading, badly constructed SQL and SQL incompatible, with current Oracle version. Various substantial estimates of requirements to complete project, based upon need for massive manpower. What would it take to run through millions of lines of code to fix issues? Conservative guesstimates were order of 40 man-years effort over elapsed time of over 1 year. In addition process had to support constant new development and on-going changes were required in code base.

Designed a process to highlight system deficiencies and fix in appropriate manner. Spent one month building tool to analyze code base. Proposed use of these analyses to highlight issues, then deploy manpower to start re-coding. After initial analysis discovered large number of patterns in "bad" code, was able to expand scope of tool to not only find code deficiencies, but fix them. Tool-set was used to make over 146,000 automated changes to code. Other issues that could not be automated were correlated for manual change, allowing all manual changes to be made in less than 80 man-hours. All changes were readily documented, minimized SQA effort, consequently total project took less than 4 months. Result was successful implementation, achieved at considerably under revised budget. After deployment massive performance increases achieved, not only in improved processing, true multi-tasking now possible because inefficient locking and transactioning issues resolved.