1. kyle mix data scientist + fraud prevention
kylermix@gmail.com www.kylemix.com 415.317.8482 San Francisco kylemix kylemix kmix27
employment
Shutterfly Inc. San Carlos, CA
Manager, BorrowLenses Fraud Dept.
Oct 2013 to Sep 2016
Responsible for all fraud prevention policies and procedures in an exceptionally high risk
industry, process owner, driver of innovation.
Implemented strategies responsible for $160k+ in prevented losses per month.
Constantly analyze past failures to implement or modify rules to combat future losses.
Maintained a constant headcount during a period of over 80% growth.
San Carlos, CA
Manager, Asset Recovery Dept.
Mar 2016 to Sep 2016
Worked closely with infosec and legal teams to ensure proper standards and compliance.
Managed a team of 2 associates responsible for over $350k in unreturned assets per year.
Personally handled claims management process when insurance companies were involved.
BorrowLenses.com (acquired by Shutterfly) San Carlos, CA
Director of Loss Prevention
May 2010 to Oct 2013
Was company's 12th hire and 1st fraud analyst, central to helping the company grow into to
a national industry leader.
Built fraud department, directly responsible for hiring team of 5 analysts and implementing
operational procedures along the way.
As comfortable with small scrappy teams as I am with large and structured teams.
projects
Movie ROI Prediction Engine
Built a tool which utilizes historic movie data to make predictions towards ROI of a future
project.
Tool helps early-stage investors with the principle goal of allowing for more informed, data
driven investment decisions, and hypothesis testing.
Key skills utilized: web scraping, API usage, data cleaning and formatting, regression
analysis.
OkCupid Classification
Explored the dataset released in early 2016, and built a tool for predicting categories of
users based on their answers to a set of 20 questions.
Focused on working with various classification algorithms and engineering features from a
large noisy set.
Web application building with flask, and visualization techniques with D3.js and Plotly.
NLP Password Strength
Utilizing Latent Semantic Indexing and n-grams of characters, built up an index of the top
35K most commonly used passwords.
Built a password scoring engine on top of that index that adds or subtracts from a baseline
entropy calculation based on a test passwords similarity to the indexed passwords, and
their relative frequency in the wild.
Built in additional metrics such as keyboard distance and letter frequency to further
manipulate entropy scoring.
Initial testing indicates an improvement at catching weak passwords over existing methods
summary
Experienced professional with comprehensive
skills in data-driven fraud detection and
mitigation, team building and management,
machine learning implementation, operations
analyses, feature engineering, and business
agility. Interested in a dynamic company to
feed my insatiable curiosity, where I can
apply my considerable creativity to
challenging and complex problems.
education
Metis Data Science Bootcamp
2016
Brooks Institute of
Photography
Bachelors of Arts Industrial/Scientific
Photography 2007
LANGUAGES: Python, Command Line, SQL,
Javascript (learning)
DATA SCIENCE TOOLBOX: Pandas / Numpy,
MySQL / SQLAlchemy, Scikit learn,
Matplotlib / Seaborn, Data Mining,
Webscraping, Mongo, Neo4J, GeoPy,
Gensim/NLTK
SOFTWARE: Jira, RightNow, Tableau,
Authorize.net, Various Fraud Platforms,
git/ github
MANAGEMENT: Team Building, Training,
Talent Retention, Grit
SOFT SKILLS: Confident Communicator,
Endlessly Curious, Appropriately Skeptical,
Creative Thinker, Gregarious
skills
activities
ACFE ·
Associate Member
Feb 2015 to
Current
For staying on top of current events and
trends within fraud prevention