@karsenthil@clairevo@gregisenberg Could be, seemed silly to me to have an agent with the same context judge another agents work, but maybe different models? Different prompts?
@karsenthil@clairevo@gregisenberg I build a lead discovery + scorer agent, run it at scale, but the system does not create many verifiable outcomes on its own
How can I shortcut that without human evaluation? I’m currently using a human eval + dspy/GEPA prompt optimization
@jacob_posel Would you be interested in an implementation partner (I.e outsourced forward deployed engineers). Working on a new concept and would love your take
@clairevo Doing this too, significantly above a 22 year old. Probably a notch below Claire!
Dm me, open to do Sunday afternoon/nights or late nights weekdays if you have a crazy packed weekdays