Previous Job
Previous
Fraud Data Scientist
Ref No.: 18-00164
Location: Pleasanton, California
Position Type:Contract
Start Date: 01/04/2018
Title: Fraud Data Scientist
Location: Pleasanton, California
Type: 6+ Month Contract

********CANNOT DO C2C*********


You will apply machine learning algorithms and statistical methods to both structured and unstructured data to minimize both false positive and false negative fraud. You'll work with product managers and engineering to build and deploy new and updated models. You will analyze model performance and communicate model design and performance results to senior management. You'll work with internal and external partners to monitor and manage fraud incidents on various products. You'll define, scope, and perform data-driven analysis and Interpret results with cross-functional teams to ensure optimal instrumentation products. You'll work with customers to help educate and mitigate card risk.
The position will be primarily located in Pleasanton, CA with option for San Francisco

Duties and Responsibilities:
  • Define, scope, and perform data-driven analysis on key strategic issues and turn business questions into actionable suggestions.
  • Support data requests for information from internal and external business partners.
  • Apply machine learning algorithms and statistical methods to both structured and unstructured data to spot trends in user behavior.
    • Use Python (Scikit-Learn, Pandas, Numpy, etc) and R to build machine learning models for both gift card acquisition and online card-not-present fraud.
    • Maintain and improve the current models in production. Promptly fix issues with the machine learning models caused by system updates, data errors, and etc.
    • Perform regular offline analysis for reporting and monitoring. For example, use Python NetworkX to build link analysis for daily reports; Use R to create summary statistics of machine learning score distribution trend over time to catch anomalies.
    • Use statistical methods to adjust fraud score thresholds on daily basis for optimized results.
    • Work with the engineering team to implement the offline developed models in production.
  • Set up and help manage a battery of AB and multivariate tests; interpret and evangelize results.
  • Work with internal and external partners to monitor and manage fraud incidents on various pre-paid products.
    • Partner within us about fraud incidents, vendor evaluation, model building process, etc.
    • Search and evaluate the performance and effectiveness of external vendors for fraud detections purpose.
    • Communicate and educate the operations team on fraud related topics. Set up guidelines/protocols for the manual reviewers and the gift card processors.


Knowledge and Skill Requirements:
  • proficient in machine learning methodologies/statistical modeling (regression, classification and unsupervised approaches)
  • Excellent written and verbal communication skills
  • Strong programming skills, with prior Python (numpy, pandas) experience
  • Experience with relational databases (e.g. SQL)
  • Experience building robust systems, or driving engineering to do so on your behalf, at scale on top of Amazon Web Services (AWS) or similar cloud services
  • Data presentation and visualization (e.g., Tableau, Python, R, etc.)
  • Experience in fraud detection
  • Ability to multi-task and work in a fast-paced environment

Qualifications:
  • B.S. required; M.S. or PhD preferred, with concentration in Computer Science, Engineering Math, Statistics, or another quantitative field
  • 3+ years of experience working with data to address key business challenges in an engineering or quantitative role, ideally with a focus on rapid feedback cycle industries (payments, gaming, social, e-commerce, online media etc.).