AI Output Evaluation and Data Labeling (NLP Tasks)
Scope: I worked on labeling and reviewing AI text outputs, mainly to help improve how the responses make sense in real situations and not just sound correct. Tasks Performed: I handled text classification, checked responses against guidelines, tagged outputs, and corrected ones that were off or didn’t really match the context. Sometimes you see responses that look okay but are actually wrong, so I had to pay attention to that. Project Size: Worked on different batches over time, going through a good number of data points consistently, not just one-off tasks. Quality Measures: I followed the guidelines closely, but also used my own judgement for edge cases. I made sure similar data was treated the same way, and I double checked tricky ones so errors don’t affect the overall output later.