At RWS TrainAI, I audit Hindi conversations for the Meta MQA ELO Chatbot Project, evaluating task success, utility, and conversational dynamics. I use a 5-Point Scoring Rubric based on the S2S ELO Guideline.
My auditing work includes identifying error clusters such as bad ASR, latency, model wrong-gender errors, inaccuracy, and coherence issues. I also check scenario adherence, scenario coherence, and rating justifications.
At Trainplex India, I train reviewers in transcription review and voice quality checks for AI models.
My background includes teaching and B2B sales work, along with a B.Sc in PCM and a B.Ed in Physical Science and Maths. I’m also a Devnagari script expert and hold RSCIT certification.
