Most recruiters are waiting for AI to hand them perfect answers.

Cassie Chao Leemans is doing something better: building the tools she wishes existed.

In this episode, Cassie, Talent Partner at Craft Ventures, breaks down how she went from “not technical” to building her own AI-powered recruiting workflows with Claude Code, Supabase, Metaview, and a lot of late-night tinkering. She shows what it actually takes to turn messy recruiting context into usable systems—and why the best AI outputs don’t come from one magical prompt.

Key takeaways:

Why “one-shotting” AI products is mostly fantasy
How Cassie built tools to prep calls, search her network, and surface high-signal candidates faster
Why recruiting expertise, taste, and judgment matter more in an AI-enabled workflow

Key topics covered:

[00:00] Intro

[01:37] Cassie builds AI tools for recruiting utility

[06:01] From recruiting to coding without technical background

[09:49] Making time for AI experimentation

[12:56] Building Radial, the AI recruiting assistant

[18:03] Iterative approach to AI product development

[22:32] Claude Code vs Claude AI decision making

[24:04] Building a custom CRM with AI

[30:07] Choosing Supabase for database architecture

[34:06] One shot AI products aren't real

[37:07] Managing broken features while building

[41:20] Trust the system you built article

[45:00] Hierarchy of hiring signal sources

[48:30] Spiky people and back channel references

[50:31] Building at the tip of AI spear

To learn more about Metaview, the AI platform for 10x recruiting teams, head to www.metaview.ai/10x