Sam Green and Aaron Linsky describe Bridgewater’s Vault, a platform for organizing internal and external unstructured research dataContext engineering designs how relevant instructions, evidence, state and tool information are assembled for a model's task. for investors and AI agents. Green uses a research query to illustrate why source coverage and time filtering matter.
Aaron Linsky explains how the team evaluated search precisionRetrieval is the process of selecting relevant stored information and returning it to an AI system for the current task. using manually reviewed examplesEvaluation is the systematic assessment of a system's behaviour or results against explicit criteria. and addressed filtering and scale for different users. The performance figures are the presenters’ reported internal results.
The presenters explain how familiar workflows such as saved searches and collections encouraged adoption. They describe an investor workbench intended to surface topics proactively and retain research findings for later use.
Watch on YouTube
