AI HR Business Partner - KI im Personalmanagement gestalten; m​/w​/d

The ORCA Group is a PE-backed buy-and-build group: eight software companies for construction and planning (CAD, cost planning, BIM collaboration, project controlling), moving from on-premise licens...

Why this role exists
The ORCA Group is a PE-backed buy-and-build group: eight software companies for construction and planning (CAD, cost planning, BIM collaboration, project controlling), moving from on-premise licenses to a subscription business. In 2026 we are replacing the commercial core systems;
Salesforce, Zendesk and Zuora are going live across the first companies this year. That gives you a rare starting position: a current stack with no legacy automation debt, and a long list of manual GTM work waiting to be automated.
You join as the first engineering hire in the new Revenue Operations team. Your job: reduce manual work across Sales, Customer Success and Support, improve data quality, and give the commercial teams more time with customers.

What you'll do

• You own automation across the commercial process from lead to churn. You find the spots where people copy, paste, type and chase, and you replace them with workflows that run, get monitored, and get used.
• AI agents and automations across the customer lifecycle: lead enrichment and routing, CRM hygiene, call and ticket summaries, support triage, onboarding and renewal workflows, forecasting support.
• Your own API integrations between Salesforce, Zendesk, Zuora and the rest of the stack; webhook- and event-driven where polling won't do.
• Evals, monitoring and cost tracking for every LLM-based workflow, so we know an automation works before Sales depends on it, and notice when it stops working.
• Monitoring and cost tracking for what you run: tracing, error handling, alerting, a clear owner per workflow.

How we work

• You report to the Revenue Operations & AI lead, who sets architecture and priorities. Inside that frame, you own your solutions.
• Stack:
  Salesforce, Zendesk, Zuora, Clay, n8n, BI tooling, Claude (API and Claude Code). You work against the LLM APIs directly.
• Rollouts are phased: pilot, buy-in, then scale. We don't push big-bang launches over the heads of the teams.

What you'll tackle in your first 90 days

• Ship two to three production automations that remove manual work for Sales, Customer Success or Support.
• Map the lead-to-churn process and identify where automation pays off most.
• Establish how automations run in production here: monitoring, error handling, data privacy, ownership.

How we measure success

• Hours of manual work automated per quarter, and whether the teams actually use what you ship
• Reliability of the workflows you run in production
• Faster lead response, cleaner CRM data, more customer-facing time for the commercial teams

After 90 days both sides can judge the fit: visible progress, measurable results, and a roadmap for the next six months.

Must-haves

• You write production code, e.g. Python and/or Type Script.
• You have built and operated your own API integrations: auth, pagination, rate limits, retries. Clearly beyond pure no-/low-code tools.
• You work with LLM APIs directly (e.g. Anthropic, OpenAI): prompt and context engineering, structured outputs, tool calling.
• You test what you build. For LLM workflows that means evals, not gut feeling.
• End-to-end mindset: you think in whole processes and take ownership from problem to running workflow.

Nice-to-haves

• Experience with agent frameworks and standards:
  Claude Agent SDK, MCP, Lang Graph or similar.
• Working knowledge of CRM/Rev Ops processes, and the Salesforce or Zendesk APIs in particular.
• Experience handling personal data under GDPR; a view on what the EU AI Act means for GTM automation is a plus.

You're a great fit if

• you'd rather ship working, maintainable workflows than write decks
• you think in end-to-end processes, not isolated tasks
• you enjoy building with code and APIs, beyond no-/low-code tools

Probably not a fit if

• you want a pure data-science or research role
• you expect fully specced tickets instead of owning problems
• you prefer low-code-only tooling

What we offer

• A GTM stack being rolled out in 2026 and real room for impact: your automations go to production, with eight operating companies as the scaling path.
• Part of a growing buy-and-build group with development perspectives.
• Office in Munich, in the countryside in Neubeuern or hybrid/part time remote (DACH/EU)
• modern equipment, learning budget

Process

• Deep-dive with the Revenue Operations & AI Lead (60 min): how you build, integrations and automations you…