Feed modulated AI

Purpose

While the power of GPT-4 is clear, its limitations have not been studied well. These limitations become very obvious when trying to build a GPT type user experience on proprietary knowledge within an organisation.

It is our conviction that this can be solved by modulating the knowledge base (aka Feed) on which the AI works and by fine tuning the model’s behaviour to specific use cases.

Approach

Organisations struggle with three phases of knowledge management

  1. Documenting: Recording inherent knowledge into artefacts that can be shared, and keeping them updated
  2. Structuring: Organising knowledge so that it is easily accessed and understood
  3. Delivering: Training or even sharing knowledge to end consumers.

This space is to record experiments in each of these three categories. Which ironically will become a knowledge base of its own.

Untitled

About us

profile pic.jpeg

RAHUL SHETTY

hhh.jpg

UDAY ARADYA

CATEGORIES OF WORK

DOCUMENTATION

STRUCTURE

DELIVERY