Data Scientist
Open Justice Lab (Contractor)·New York, NY
- Implementing causal inference methods to identify variables driving deaths in custody and recidivism.
- Processing and cleaning data produced by Boston University Spark! teams.
- Running self-hosted vLLMs to convert FOIA-gathered PDFs into usable tabular data via OCR.
- Submitting and managing FOIA requests for state prison capacities, prison building materials, deaths in custody, and gun deaths by county.
- Merging BJS census snapshots that lack shared facility IDs using probabilistic record matching.
- Using Bayesian sampling and modeling to estimate per-document error rates for OCR-converted records.
- Building websites and visualizations so litigators, researchers, and lay readers can explore the data directly.