Recruitment 02
Xoots.ai
A from-scratch frontend that unified recruitment workflows and supported 100+ hires across client organizations.
My contribution
I built the frontend from scratch across Interview-X, Talent-X, Academy-X, and Search-X, choosing SSR or TanStack Query based on each workflow’s rendering and server-state needs.
Shipped result
The platform supported 100+ hires across client organizations, while Captivate Chat also used Xoots internally to identify and hire top talent.
01 Context
AI hiring tools can fragment the workflow and obscure how a recommendation was reached.
The suite supports recruiters and candidates across discovery, interviewing, assessment, and learning. The interface has to communicate progress and confidence without presenting AI output as unquestionable fact.
Working within
- Long-running and failure-prone AI tasks
- Rendering strategy across public and authenticated workflows
- Multiple products sharing language and state patterns
02 Product decisions
AI output is a draft, not a verdict
Recommendations remain inspectable and reviewable, with source context and clear human actions nearby.
Match the data strategy to the workflow
SSR supports server-rendered entry paths, while TanStack Query manages caching, refetching, and interactive server state where the product needs it.
One suite, different levels of trust
Shared primitives preserve coherence while interview, search, and learning flows expose different evidence.
03 Reflection