Resume

Ugur Yildirim

LinkedIn: https://www.linkedin.com/in/ugur-yildirim-soc

Medium: https://medium.com/@uguryi

GitHub: https://github.com/uguryi

Twitter: https://twitter.com/UgurYildirimSoc

Education

University of California, Berkeley
Ph.D. in Sociology

2014 – 2020

University of California, Berkeley
M.A. in Statistics

  • Relevant coursework: machine learning, natural language processing, probability theory, statistical computing, generalized linear modeling

2018 – 2019

Bogazici University
B.A. in Political Science and International Relations
B.A. in Sociology

2009 – 2014

Skills

  • Programming: Python (Numpy, Pandas, Matplotlib, Seaborn, Scikit-Learn, Statsmodels, SciPy, TensorFlow, Keras, BeautifulSoup, Requests, NLTK, PySpark), R (dplyr, ggplot2, tidyr, tidyverse, car, glmnet, MASS), SQL (Presto, Hive, SparkSQL)
  • Methodologies: Experimentation (A/B testing), causal inference, machine learning, time series, natural language processing, network analysis, contextual bandits
  • Tools: Jupyter, Looker, Airflow, UNIX, Git, Stata, Qualtrics, MTurk, LaTeX, Breadboard

Work Experience

Slingshot Finance – Data Scientist

  • Built the company’s incentive strategy from the ground up, ran numerous promotions that contributed to more than 2x increase in user base and revenue
  • Developed a graph model to identify and remove fraudulent wallets in our incentives, which helped save cost by more than 50%
  • Created and maintained Looker dashboards for key metrics, which were used by the CEO to track app performance, and results were shared in All-Hands
  • Built an MAU projection model based on company-specific milestones, presented it to the investors, and helped secure millions of dollars of funding

May 2022 – Oct 2023
San Francisco, CA

Robinhood – Data Scientist

  • Designed and executed incentive experiments that brought in 10,000+ customers in a cost-efficient manner assessed in terms of incremental payback period
  • Productionized a time series model to predict trading volume as a function of external factors, built an Airflow workflow with Looker components to visualize and monitor the results daily, and findings were used as an input for the Earnings Call
  • Developed a Python package to estimate causal impact using observational time series data, which was used to analyze the performance of multiple promotions
  • Mentored a new hire data scientist on my team, designed their starter project related to standardizing the promotion funnel conversion queries, guided them through the successful completion and presentation of their project

Apr 2021 – Apr 2022
Menlo Park, CA

Quora – Data Science Intern

  • Analyzed historical data using Python and SQL to assess the impact of slow processing times and long backlogs at the question topics queue
  • Initiated the optimization of the question topics queue by safely removing stale questions, which saved the company around $400,000 for the fiscal year 2019

Jun 2019 – Aug 2019
Mountain View, CA

Thusly – Computational Social Scientist

  • Scraped and parsed an online text corpus of thousands of articles using Python, which was used on the company’s data labeling platform
  • Built a custom named-entity recognition classifier using Python and Stanford CoreNLP

Oct 2018 – May 2019
Berkeley, CA

Social Sciences D-Lab – Consultant

  • Guided clients through statistics and programming (Python, R, Stata) related questions, including machine learning basics, co-occurrence analysis, dimensionality reduction (PCA), web scraping, and GitHub

Jan 2018 – May 2018
Berkeley, CA

Research Experience

2020

Implemented the ABE method for historical record linking as an R package (abeR) and applied it to the CenSoc mortality datasets

2017 – 2018

Drafted survey questions and built weighted percentage tables in R for the Bay Area Poverty Survey project

2017 – 2018

Authored a series of Stata files that cleans and joins company datasets using regular expressions and exact/fuzzy matching for the Shift project

2015 – 2016

Reviewed the literature and prepared percent stacked bar charts and other figures for the book project A Fraught Embrace: The Romance and Reality of AIDS Altruism in Africa

Selected Publications

Yildirim, Ugur. 2024. “An Overview of Contextual Bandits: A dynamic approach to treatment personalization.” Towards Data Science Editors’ Picks.

Yildirim, Ugur. 2024. “Sensitivity Analysis for Unobserved Confounding: How to know the unknowable in observational studies.” Towards Data Science Editors’ Picks.

Wodtke, Geoffrey T., Ugur Yildirim, David J. Harding, and Felix Elwert. 2023. “Are Neighborhood Effects Explained by Differences in School Quality?” American Journal of Sociology 128(5).

Yildirim, Ugur. 2020. “Disparate impact pandemic framing decreases public concern for health consequences.” PLOS ONE 15(12): e0243599.

Connor, Paul R., Daniel Stancato, Ugur Yildirim, Seunghun Lee, and Serena Chen. 2020. “Inequality in the Minimal Group Paradigm: How Relative Wealth and its Justification Influence Ingroup Bias.” Journal of Experimental Soc. Psychology 88: Article 103967.

Teaching

2018

Demog 180, Social Networks (Instructor)

University of California, Berkeley

2016, 2017

Soc 5, Evaluation of Evidence (TA)

University of California, Berkeley

2016

Soc 1, Introduction to Sociology (TA)

University of California, Berkeley

Selected Honors and Awards

2019 – 2020

Won three separate research grants from the Experimental Social Science Laboratory (XLab) at UC Berkeley to run two survey experiments and one behavioral experiment as part of my dissertation research