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:
[
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