Joseph Baum Email & Phone Number

Boulder, Co, Us

Des Plaines, Il, Us

San Francisco, California, Us

Data scientist supporting DoorDash New Verticals business (i.e. grocery, convenience, and retail store deliveries)

Los Angeles, California, Us

Data scientist/tech-lead for e-commerce and catalog listing ad solutions

Mountain View, Ca, Us

- Supporting Google Search data science and eng infra

- Analyzed advertiser lifecycle behavior that resulted in identifying a key asymmetry in advertiser churn and resurrection behavior- Contributed to the development of the AB testing infrastructure; contributed to the development of Facebook's synthetic control feature and techniques to handle various pre-bias issues (e.g. regression adjustment, stratified sampling, etc.)- Improved new advertiser retention by leveraging various ML techniques which were guided by a combination of data deep dives and insights from XFN UX qualitative research

San Francisco, Ca, Us

- Built and maintained dashboards to track various KPIs to analyze the success of various product feature launches- Designed and evaluated experiments (e.g. A/B tests)i) Developed a Bayesian based A/B testing framework- Helped develop a user behavioral model to identify key driving factors of user growth and retention

Fremont, Ca, Us

- Developed a business intelligence tool that performed ETL tasks and data analytics on the terabytes (>100 TBs) of data being generated at Seagate’s overseas manufacturing HDD facilities- ETL processing performed using Python and SQL scripts (executed and created in real-time using Python) i) Data sources: multiple (overseas) Oracle DBs, binary files, csv files, and log (unstructured text) files ii) Data storage: (local) MySQL server (using the HeidiSQL and MySQL Workbench IDEs to manage the server)- Data analytics and visualization was performed using R and Python

Fremont, Ca, Us

- Analyzed trends in (consumer) data usage behavior to aid in the development of Seagate’s hybrid HDD program i) Data source: binary (unstructured) log files generated from the HDD SATA drive interface ii) File sizes: 2-10 GB per usage case iii) Data analytics and visualizations were done in R- Identified key marketing advantages to help develop the hybrid drive line i) Findings were presented (by me) to the on-site VP and head of marketing ii) Results and key visualizations were developed into marketing collateral used at trade shows and were included in ExtremeTech's (3rd party) review of the hybrid drive (*see link below)- Data analytics performed (on the consumer data sets) identified key patterns in usage behavior and led to the development of a superior file-caching (patent approved) algorithm i) Patent Title : Cache Data Value Tracking ii) Patent Number: 9,213,646

New York, Ny, Us

- Analyzed StubHub customer purchase data (50,000+ observations) to model and forecast customer lifetime value- Constructed an integrated model (customer dropout process and transaction process) via a hidden Markov model- Dropout/Transaction process – modeled using a shifted beta-geometric and beta-binomial distribution

- Imported environmentally friendly bags from China and Singapore for wholesale distribution- Customers included both individuals and businesses from Canada and the United States- Customers included: the ‘BolderBoulder’, the Denver Broncos, and Western Union

Bachelor’S Degree, Finance & Biochemistry (Vagelos Life Science Management Program)

Phd (Dropout), Applied Mathematics -Statistics/Probabilistic Modeling (Emphasis On Hierarchical Bayes)