All projects

Counselling summary and diarization

From raw audio to a per-speaker counselling summary

TebahSoftMar 2023 – Jul 20235 mos

Context

Writing up session notes ate directly into counsellors’ working hours.

Problem

Counsellors wrote up their notes by hand after long sessions, spending real time sifting out what each speaker actually said.

How I solved it

  1. Designed an end-to-end pipeline: speaker diarization, then speech-to-text, then LLM summarisation
  2. Delivered per-speaker summaries and a whole-session summary as separate outputs

Result

  • Audio in, per-speaker and full summaries out — and it shipped to production

Impact

Audio in, per-speaker notes out — counsellors spend their time counselling rather than transcribing.

Tech

PythonPyTorchStreamlit
NextSabeujakLaunched