
Data Annotator Intern (NLP Analytics)
Prodigal Tech · Sep 2025 – Present
- Analyzed 600+ Al-generated NLP outputs weekly against human benchmarks, categorizing errors and quantifying error rates to improve data accuracy for ML teams.
- Analyzed edit-rate and note-volume trends across 16 client accounts (up to 190,000 notes/day for the largest), identifying the top error drivers used to prioritize model retraining.
- Isolated low-headcount accounts that skewed aggregate error rates, enabling more accurate account-level performance reviews.
- Built and maintained reporting pipelines and visualizations (graphs, tables, charts) that communicated performance trends to management.
- Partnered with ML engineers to turn data findings into recommendations, and cleansed and normalized annotated datasets across multiple clients.

