
Fine-Tuning vs Model-Agnostic Approach

Ex-Google Duplex Eng Lead on Disrupting $2B Clinical Trials with AI - Patrick Leung
Topic 7 of 18
Published Jul 28, 2025 · Clip 16:41–18:50 (2:10)
Patrick explains that Faro hasn't needed to fine-tune models because major vendors like Google and Anthropic focus on life sciences, and Faro's system is model-agnostic—their evaluation system lets them easily assess which LLM produces the best clinical output. They may build their own vectorization scheme for clinical terminology in their RAG system, but so far evaluation plus prompt tuning suffices to produce clinical-grade documentation.