Tiyasha Das

I’m an M.A. Linguist with hands-on experience in phonetic transcription, corpus annotation, and linguistic data analysis. I work across Bengali, English, and Hindi, using IPA and standard phonetic schemas to turn raw language into dependable, research-ready data.

In my phonetics and language documentation work, I annotated and transcribed 250+ audio samples with 95% accuracy, and performed structured text annotation on a 250 corpus using consistent labeling guidelines to support inter-rater quality standards. I also documented 50 phonological patterns across Hindi speech data and translated findings into structured analytical reports.

I bring that same linguistic rigor to AI by evaluating and rating AI-generated responses for quality, accuracy, and language naturalness as part of an LLM training and RLHF data pipeline. I’m excited to apply deep linguistic knowledge to data annotation, AI model evaluation, and LLM training pipelines.