Zach Ellis

AI Research Engineer - Ufonia

[Zach Ellis @ZachEllisX]

Excited to share our work on improving safety evaluation of Clinical Voice AI transcription, using @DSPyOSS and GEPA to build a clinician-level LLM Judge!

[Jared Joselowitz @JaredJoselowitz]

🚨 ASR errors in clinical dialogue can be dangerous, and WER doesn’t know it. Today we release “WER is Unaware”.

Using DSPy + GEPA, we optimise an LLM Judge that reaches clinician-level performance at detecting safety risks.

🔗 https://t.co/RZOKEa4JWq
📄 https://t.co/h5YxKghv3W