GrokATS [Recruiter Track]

• 

  Source candidates by dispatching traversal agents. Specify how large you want the tree

• 

  Collect incredibly detailed data

• 

  send out automated AI screens (results used to update agents' memory)

Before AI, talent was a social problem.

Now, it's an engineering problem.

We send out AI agents to traverse the X social network, following threads and tapping into the technical community and the conversations therein. We also scour github and arxiv for eligible candidates. The whole operation is organized into a tree -- this keeps the agents on a defined (and unique) track, minimizing redundancy. It also allows us to parallelize to any extent.

We can automate outreach to candidates, simply sending them a number (that they can call at their convenience) to do an AI recruiter screen. We also engineered an automatic technical system design interview, which uses a Video agent to collaborate on a canvas, providing feedback, and asking follow up questions.

All this data is collected across the recruiting stack and used as memory for the agents. For example, if an agent finds someone online, it will check this memory to see if: the person has been interviewed before, the person has red flags seen in other candidates, or the person has strong green flags seen before in successful hires.

The end goal of this platform is to replace all stages of the hiring process before hiring manager. The HM will solely use this platform to source, screen, and decide on candidates -- no other humans needed. And with constant yes or no feedback on the candidates, we intend for the platform's agents to be getting more and more accurate.

Built With

• grok