Transcription Fraud Analysis and Voice AI Evaluation – JoshTalks PM Internship Assignment
As part of the JoshTalks PM Internship Assignment, I analyzed transcription fraud within a dataset of 51,783 submissions. I identified fraudulent patterns and recommended detection logic for AI-powered content evaluation. I also designed a Voice AI evaluation system, detailing quality metrics and testing frameworks. • Evaluated text submissions for transcription accuracy and authenticity • Identified and catalogued fraud patterns within large textual datasets • Recommended detection logic to enhance AI fraud detection capabilities • Defined and structured quality metrics for Voice AI assessment