Loading views...

HAI Agents Planning

Creator
Creator
Seonglae ChoSeonglae Cho
Created
Created
2025 Dec 16 13:31
Editor
Edited
Edited
2026 Jan 23 18:5
Refs
Refs
definition and scope of the task - qualitative results or final output from holistic ai side not intermediate
dataset
model
custom api
documentation or paper to show client

Three pillars

  • llm
    • custom assessment
  • agent
    • agent graph
    • custom assessment on state graph layer
  • robotics
    • custom robotics
human in the loop.

Core philosophy

  • we need to let user to spend more time on the platform
  • for that we should make a sustainable, reliable, and customisable platform such as app store including developer and governance team
  • What we need to do specifically
    • Refactor the codebase to reduce buggy features
    • Make customizable features based on 3 layers
  • By doing these two, we can build sustainable platform which give us a time to develop further features and research such as robotics

Big thing

  • tracker
    • rishi
    • human in a loop legal anisha
  • task execution agent (asset agent/ notification)
    • human in a loop
    • umar
  • refactoring / unit test
    • no more features
    • seonglae
  • multi tenancy / UAM
    • kleyton
  • Agent tracing
    • ilham
  • Clustering
    • christian
generalization → specialization
main issues → sub required tasks
 
Big problems what we are facing now
main issue: our platform is not ready
 

Direction

  • maturity
  • reliability
  • agent with human in a loop
 
 
 
 
 

Recommendations