AI with Michal
Recruiting Systems Build · 3–4 weeks

Internal Talent Pool Infrastructure

Stop losing past silver-medalists. A searchable, structured talent pool that pays back the next time you open the same role.

  • Every candidate from the last 12–24 months indexed in one structured database
  • AI-powered semantic search ('engineers who built event-driven payments backends in EU')
  • Re-engagement workflows that surface latent fits the next time a role opens
  • Tagging + freshness signals so the pool stays usable as it grows

Starting at €5,000 · 3–4 weeks

Request this build

Scope

What we build

In scope

  • Audit of your current candidate data (ATS export, sourcing sheets, LinkedIn outreach history)
  • Database design (Notion/Airtable/Postgres — your call) and import pipeline
  • Semantic search via embeddings + Claude reasoning layer
  • Re-engagement workflow templates (cold-start, warm-touch, role-match)
  • Owner handoff session + freshness/maintenance plan

Out of scope (separate engagement)

  • ATS replacement (we layer alongside, never replace)
  • GDPR consent re-collection campaigns (we'll flag where this is needed)
  • Multi-tenant deployment across separate business units

Deliverables

What you get

  • Talent pool database

    Production-ready Notion / Airtable / Postgres database with all historical candidates imported.

  • Search interface

    Semantic search via embeddings + Claude — usable from a simple web form or your existing tool.

  • Re-engagement workflows

    3 templated workflows (cold-start, warm-touch, role-match) with prompts and runbooks.

  • Maintenance plan

    How to keep the pool fresh, what to tag, when to prune — written + recorded.

Is this the right fit?

Best fit when you have 500+ past candidates lying in spreadsheets, ATS exports, or LinkedIn DMs.
Smaller teams typically don't have enough data yet to justify the infrastructure.

Ready to start

Request a Internal Talent Pool Infrastructure

Tell us the outcome and the constraints. Two-business-day response. If it isn’t a fit, we’ll say so and recommend the better path.

More from the catalog

Other builds