VP, Analytics & Modeling - American Home Mortgage Servicing (Coppell, TX)

Statistical modeling/Analytics Strategy will provide leadership to a team of model developers. This hands-on manager will lead, manage, and execute highly complex predictive modeling initiatives that are critical to the firm's success - strategic, tactical, & operational; whole loan acquisition pricing, mortgage servicing portfolio management, consumer behavior analysis, Home price trends, macro and micro economic trends, MSR valuation, credit risk, interest rate risk, etc. The professional in this role will establish/maintain relationships with all levels of management, initiate, prioritize, scope, design, and deliver high-visibility quantitative projects. Provide assistance in translation of business needs into technical requirements and insure high-quality and timely delivery of results.

Manage and provide leadership and direction to a team of quantitative analysts (including international teams)

• Research and deliver new ways of leveraging our extensive mortgage data infrastructure and provide data-driven insights.
• Collaborate with key stake holders across various levels in the organization and serve as champion of analytics
• Responsible for planning, organizing and leading tasks and projects, and insuring the quality of the deliverables.
• Respond to the needs of management and external entities for data, analytical and quantitative support
• Other duties and responsibilities as assigned.

Qualifications

• Graduate degree (Ph.D. preferred) in a highly quantitative field : Statistics / Bio-statistics / Computer Science with machine learning/data mining focus. Strong interest in delivering practical solutions that will have a quantifiable contribution to the business.
• Minimum of 5+ years of experience in modeling/data intensive environments, providing competitive advantage/arbitrage using cutting edge analytics.
• Expertise: Practical and theoretical, in a broad spectrum of analytical (modeling) techniques:
  • Least squares, maximum likelihood, GLM, time series, SEM, mixed models, survival models, panel regression, competing risk.
  • modern techniques such as neural nets, Kohonen, SVM, Classification trees, EM,
  • Meta learning/regularization techniques such as boosting, bagging, Lasso, etc.excellent grasp of EDA/data transformations
  • Experience in handling large, complex and messy data and expertise in key data sources used in mortgage industry.
• Substantive experience working with senior business managers
• Intellectual curiosity and honesty + Ability to package complex solutions in the business context
• Strong verbal and written communication skills
• Strong strategic, analytical, and creative thinking aptitude in both individual and team environment
  • - Holds self and others accountable for timely, high-quality results
• Experience in risk mangement environments is preferred (credit risk scoring, PD,LGD, delinquency/prepayment predictions, HPI
• Software: SAS (base,stats,ETS,enterprise miner, macros)/R etc., Good understanding of database/sql/software principles and code structure.

About AHMSI, American Home Mortgage Servicing, Inc.:

Established in April 2008, AHMSI brought together two mortgage loan servicing platforms, each with strong capabilities and specialized expertise. By multiplying these strengths, the company has achieved more than the sum of its parts – it has taken servicing to new levels. With a multibillion portfolio under management, AHMSI is one of the country’s largest servicers of Alt-A and subprime loans.

AHMSI is based in Coppell, Texas, with servicing operations in Irvine, Calif., Jacksonville, Fla., and Pune, India. It is funded by Wilbur Ross & Co. LLC., a private equity firm based in New York, and holds membership in the Mortgage Bankers Association and the HOPE NOW Alliance. Differentiated by a focus on creative solutions and innovative ideas, the company credits its success to a talented team who do the best possible job for its customers.

To apply: click on http://www.datashaping.com/jobs18379x.shtml