Candidate pipeline (LeverTRM)
LeverTRM merges ATS and CRM in one view, which is the selling point — until you hit the customisation walls and the performance dips as the database grows.
What people actually say
✕Handling large talent pools and archived candidates becomes cumbersome at scale, and search/filtering noticeably underperforms once the database is sizable.
Source: Skima, Lever deep-dive review ↗✕Reviewers describe recurring bugs, slow support responses and unresolved technical issues, with UI responsiveness that 'could be snappier' as more features layered on.
Source: Capterra, Lever reviews ↗✕Boolean candidate search struggles at scale, and integrations have a habit of breaking after platform updates and taking time to get fixed.
Source: Treegarden, Lever review 2026 ↗
We model the pipeline as a Postgres-backed state machine — stages are typed transitions per requisition, not a one-size template. Candidate search runs on Postgres full-text plus tsvector + pg_trgm (and Typesense if you need ranked relevance on millions of rows), so it stays fast at 200k archived candidates. No per-headcount tax for inviting hiring managers as commenters.