AI/LLM Analyst at Innodata (LLM dataset annotation and quality assurance)
Performed LLM dataset annotation by synthesizing high-volume datasets through annotating and categorizing text and multi-step reasoning traces per technical guidelines. Applied predefined rubrics and taxonomies to generate structured, high-quality outputs intended for LLM benchmarking and model alignment. Flagged inconsistencies, ambiguities, and errors in reasoning chains to ensure dataset integrity and improve AI system performance. • Annotated reasoning traces and categorized content according to detailed guidelines • Used strong reading comprehension to analyze complex agent actions • Provided structured feedback for model refinement • Ensured quality via error detection and consistency checks