Igor Geyn, PhD
Applied Data Scientist
igorgeyn@gmail.com · LinkedIn · SF Bay Area
Applied data scientist and researcher with publication record and 10 years of experience. Specialize in clean delivery of complicated data analysis, workflow automation, applied machine learning (ML), and causal inference using both experimental and observational methods (e.g., A/B testing, RCTs, RD, and DiD). Comfortable in fast-paced, rigorous settings where projects span across business functions. Recognized for professional excellence and leadership skills.
Experience
Held during PhD program
Conduct research on workforce trends as part of a team of economists. Leverage public and proprietary data to answer policy and business questions (e.g., designing workforce development programs) for clients. Provide subject matter expertise on research design, statistics, causal inference, and ML. Build user-facing dashboards and other visual outputs.
Automated analysis across dozens of enterprise risks (billions of USD in expenditures) using Python, SQL, and AWS. Reduced preparation time, increased process transparency, and shortened regulatory filings by weeks. Built interactive tools (e.g., dashboards) for exploring risk scenarios, calculating program effectiveness, causal analysis, and communication to senior stakeholders. Performed first causal analysis of wildfire prevention programs (a key PG&E risk exposure that led to company bankruptcy in 2019). Promoted from Data Science Intern.
Combined causal inference techniques (geo discontinuity, SynthDiD) with re-weighted customer survey data (MrP) to plan product expansion (R + Python). Analyzed dataset with thousands of user survey responses and millions of rides generated by in-app customer behavior. Executed several natural experiments to identify previously unseen business opportunities.
Co-authored 4 policy reports on D.C. health workforce access and disparities. Combined ACS data, FOIA records, and original geocoding to produce first publicly available analysis of D.C.'s health workforce.
Evaluated California health program effectiveness using difference-in-differences methodology.
Wrote production code to enhance customer data. Conducted UX research using ML to improve contract detail prediction and linkages across millions of contracts. Improved web application experience for BGOV and Bloomberg Terminal products. Assisted company-wide cloud migration.
(2x) Led multi-analyst team in execution of multi-million dollar flagship research product delivered to Fortune 500 companies. Ensured timely delivery of complicated contracts and nuanced market/marketing analyses using MaxDiff, conjoint, and other approaches. Fielded surveys to 250,000+ respondent audience using complex weighting and respondent quota structure. Performed survey and white label research for 100+ client engagements using R and Qualtrics. Promoted from Research Analyst.
Conducted quantitative and qualitative research for grant-funded projects on health workforce shortage, regional disparities, pilot program evaluation, and forecasting (primarily R). Co-authored 4 policy reports including statewide physician workforce analysis for the California Health Care Foundation. Coordinated large, state-funded evaluation programs and other projects.
Education
Technical Skills
Honors & Awards
Some More About Me
My dissertation uses real-world, non-lab data and causal research methods (e.g., regression discontinuity design and diff in diff) to provide concrete answers to long-standing questions:
- Does attending college make people more active citizens? (Yes, by quite a bit.)
- Does direct democracy create a more informed electorate? (It depends on how you measure it, but probably not.)
- Is the administration of U.S. elections fair or biased in a partisan way? (Fair, but it matters who is in office.)
I also have experience teaching statistics and data science/research methods, both formally (through 3+ years of teaching at UCLA) and as a tutor/coach for college students and young working professionals.
I've been lucky enough to develop my data skillset alongside growth in the sophistication of data analysis overall, having worked in every type of infrastructure/computing environment from scattered Excel workbooks to robust, integrated cloud pipelines. This has required me to learn a lot more than just coding and analysis: stakeholder management, task scoping, people management, and generally how to make do with what you have while working/building towards a better future state. Some examples of my project-level achievements follow.
- Managing a team of researchers to execute the flagship research product, Leading Brands in Government, at the public sector division of Atlantic Media.
- Piloting and executing the first explicitly causal analysis of wildfire prevention programs at Pacific Gas & Electric Co. (PG&E).
- Launching the first political science program at UCLA to teach undergraduates research skills by pairing them with high-quality, in-flight graduate research projects and teaching modern-day research methods.