Data Annotation Specialist
I worked on a large-scale project focused on extracting key data from financial news, company reports, and SEC filings. My main responsibility was using Named Entity Recognition (NER) to precisely tag entities like corporate names, stock tickers, and monetary values to help improve a market analysis model. The work required a high level of focus to maintain a 98% accuracy rate, especially when dealing with complex text where I had to distinguish between parent companies and their subsidiaries. I also spent time reviewing "edge cases" where the automated system struggled, manually correcting labels to ensure the training data was top quality. By the end of the project, our team’s annotations directly contributed to a significant increase in the model’s ability to pull accurate data from messy financial documents.