
President Multivariate Diagnostics LLC
Greater New York City Area

President Multivariate Diagnostics LLC
Greater New York City Area
PROFESSIONAL SUMMARY- Responsible for database optimization and systems development that lead to process re-engineering, and then reporting of key non traditional business metrics and insights back to senior level management. Defined program strategy, vision, and infrastructure requirements that internally integrates these business objectives and supported data driven decisions by using my experience synthesizing disparate data and applying multivariate techniques to uncover leveragable opportunities while focused on systems delivery using analytical dashboards and visualizations. Previous SAS invited forum speaker on data mining and current SAS author with book entitled ‘Data Mining with SAS: Multivariate Functions and Segments’ in progress.
An experienced business intelligence statistician within the quantitative decision support industry highly skilled with SAS Enterprise Miner, Enterprise Guide,multivariate predictive classification techniques, ARIMA time series approaches, CHAID/Tree decomposition, neural algorithms, and creating variable transformations from optimized binning. Data sources include large transactional files, enhanced client master files and financial metrics. Able to translate client concerns into deliverables that provide immediate impact using excellent communication and interpersonal skills to provide deep dive working end to end solutions and systems.
Statistical Toolkit - SAS 9.2, EMiner 5.3, ODS, WEB Report Studio, Proc SQL with ODBC, COGNOS 7,
SAS 9.1.3 SP4 , EGuide4.0, SAS Management Console 9.1, StatStudio3.1,SimulationStudio1 ,XML Mappper, ODS, Graphic Language Template , OLAP and BI.
Personal Toolkit - Disruptive thought leadership about leveraging business analytics, predictive statistical models.
The art and science of mathematical investigation begins with asking, “What kind of data do I have? “, then “How do I re-represent these raw numbers into something more meaningful?” Datum refers to a single metric, usually a number while data are a collection of numbers and alpha-numerics. Before selecting a tool and building models, the investigator needs to create an extended database from which to calculate their models.
(Management Consulting industry)
2005 — Present (4 years )
How to provide price discovery for index and deal level RMBS ?
Designed an all optic Fiber Optic RAID with enhance performance workstation configuration using SAS 9.1.3. Able to provide 500-750 MB/sec transfer rates with added performances from SPDE SAS engines. Housed and supported a 1B record DB (4TB) with all reporting and analytics on a loan level basis performing cross correlational pricing volitility on individual deals and broader market indexes. Data sourced from Corelogic – Loan Performance
How to identify driving factors within a rating agencies dollar loss severity model
On a loan level basis , modeled the expected dollar loss reported by a rating agency against the complete credit agencies records on that loan. Provided a redistribution across tranches earning additional bips on $800MM in deals privately placed through Bear Stearns.
Issue – How to leverage credit card transactional files into new revenue producing products ?
Created informational products for Citigroup based upon credit card transactions with focus on the consumer sector profiling, government statistic predictions and equity market performance.
Details – Multivariate Algorithms Inc wholly owned by Daniel Kocis, Ph.D. provided his personal supervision over 6 supplied analysts that advanced the Franchise Leverage Division at Citibank NA 2006-2008. One effort focused on market share and competitive dominance for the US Hospitality industry. A concurrent effort focused on translating 65M daily credit card transactions into specific equity valuations and various government statistics ( Real-time GPD estimates as Citi transcations represented 5% of total GDP) and inferring non-farm unemployment from spending and paying patterns.
(Medical Devices industry)
January 2004 — Present (5 years 11 months)
REALTROMINS:Advanced decision support for assessment and tracking the risk of mortality
The art and science of bioinformatic investigation begins with asking, “What kind of data do I have? “, then “How do I re-represent these raw numbers into something more meaningful?” Datum refers to a single metric, while data are a collection of numbers with time dimension transformation allowing for tracking dynamic shifts through out the patient stay. However, unorganized data obscures the view of the patient as a whole, hiding key facts that deserve attention, and complicates the delivery of relevant electronic knowledge to improve decisions. An “optimized binning and time-transformation” approach was developed to focus attention onto those specific ranges found within a variables total range of values.
REALTROMINS is a “new” bioinformatic system technology originally created in the Pediatric ICU (PICU) at UNC to continuously assess and warn bedside care providers about the changing status of critically ill pediatric patients and tracks their responses to ongoing medical interventions every two minutes. Bedside staff can determine which therapies are decreasing the mortality scores, which would allow for weaning of risky interventions. Alternatively, an increased mortality score will trigger new approaches to the clinical problem. Patient outcomes improve by displaying the overall risk score and the individual components, such that, contributors to that level of risk are identified and appropriate interventions applied.
In summary we have used advanced signal processing, data transformations, optimized binning, statistical data fusion, and non-linear data mining techniques to create a real time, continuous, composite index of risk for mortality and/or instability (REALTROMINS- Real Time Risk Of Mortality and Instability) in critically ill children. (NIH Grant 1 R41 HD047687-01 Multivariate Algorithms Inc 2005)
(Management Consulting industry)
November 1996 — Present (13 years 1 month)
FFIEC had no guidelines of market risk based potential for federal regulators
Housed and modeled a credit agencies master file of over 300 STAG and demographics for each individual in US, creating segmentation models targeting mortgages, refinance-home equity, credit cards, personal, and auto loans. Testified before congressional banking sub-committee and sold business to credit agency.
Details – Financial Modeling Concepts had a joint venture with Bear Stearn for CRA Securization and partially owned by Trans Union.
Co-Founder, Partner and Chief Scientist – Financial Modeling Concepts 1994 - 1998
How to leverage credit card transactional files into new revenue producing products ?
Created informational products for Citigroup based upon credit card transactions with focus on the consumer sector profiling, government statistic predictions and equity market performance.
How to compare operating expenses on commercial properties by size and location ?
Created a monthly cost accounting system that supplied online reporting and visualization tools that tracked property level expenses across a standardized chart of accounts and incorporated BOMA averages
How to establish an informational conduit to assess risk based acquisition and retention
Engineered a working Enterprise Reporting System that exploded information throughout a major credit card bank by alligning default and balance transfers against new marketing efforts.
Details - Implemented quantitative selection models for balance transfer credit card solicitations at PeoplesBank and provided in-house training and resources that established an 25 member quantitative modeling department. Won this multi-year contract over McKinsey.
(Public Company; Financial Services industry)
2006 — December 2008 (2 years )
How to cut total advertising without impacting call center traffic
Using inclusionary and exclusionary profiles established a finance risk reporting system that drilled down on Ad effectiveness and ROI which focused on driving conforming mortgages capable of earning par plus at time of securization with GSA.
Details – Multivariate Diagnostics LLC (wholly owned by Daniel Kocis Ph.D.) reporting to John Bunnel CFO of GMAC RESCAP (dba ditech ) 2006-2009
Developed, analyzed and reported dashboards from a risk base enterprise system on a 250TB Oracle database located in Texas, analyzed with business intelligence software located in Minnesota, reporting through headquartes in Costa Mesa CA, maintained by Multivariate via VPN in New York.
(Public Company; Financial Services industry)
2006 — 2008 (2 years )
See ditech
(Management Consulting industry)
2003 — 2007 (4 years )
(Online Media industry)
2004 — 2006 (2 years )
Managed web analytics for ING and John Kerry’s Presidential bid. ING was a branded site co-marketed to AOL subscribers using banner-ads served by Atlas-DMT. JohnKerry.com used OMD banner ads to drive online contributions with weekly adjustments to the media buy based upon contributions collected.
Managed all web analytics for Abbott Labs that used unbranded sites (FightRA.com and RA.com) as teasers to redirect qualified leads to Humira.com through the use of direct TV ad.
(Financial Services industry)
March 1985 — January 1990 (4 years 11 months)
: VP – Technical Director at Citicorp. Quantitative marketing of Insurance products to card members, created new acquisitions models for Private Bank, compiled and edited “BanannaBook” – A Financial Share-of market Annual Evaluation using the USCB (the original Single source householded marketing DB across all products). Experienced with the monthly card transaction DB (240MM).
1976 — 1980
1966 — 1969