Prompt Engineer.

Extracts reliable, repeatable quality from existing AI tools.

Remote & on-site DACH-wide
Prompt Engineer — enquire

A Prompt Engineer designs, tests and systematises the instructions that language models work with — from prompt libraries for business units and system prompts in products through to few-shot strategies for complex tasks. The goal is repeatability: the same task, reliably delivering the same quality.

The role is often underestimated and often overestimated. Underestimated, because good prompts in products make the difference between 70 and 95 per cent usability. Overestimated, because prompting alone does not solve architectural problems — which is why good Prompt Engineers work closely with LLM and RAG engineers.

01

Prompt library for business units

Sales, HR or Procurement receive tested, documented prompts for recurring tasks — instead of everyone experimenting on their own.

02

System prompts for products

The invisible part of your AI feature: tone, boundaries, format fidelity and escalation behaviour, tested against eval datasets.

03

Stabilising output quality

An AI workflow sometimes delivers brilliantly, sometimes uselessly. Systematic prompt iteration with measurement instead of trial and error.

04

Employee enablement

Hands-on training that goes beyond "write politely with context": role-specific patterns for real work tasks.

Core competencies
Prompt patternsEval-driven iterationFew-shot designOutput structuring (JSON/format)Instructional design
Tools & frameworks
Promptfoo/eval toolsOpenAI/Anthropic/GeminiPrompt versioningA/B comparisons
Plus factors
Domain expertiseMultimodal promptsAssessing the transition to fine-tuning
How do you test whether prompt A is better than prompt B?
The watershed of the role: measurable evaluation against test cases vs. "feels better".
When can a problem no longer be solved with prompting?
Credibility: good Prompt Engineers know the boundary to RAG, fine-tuning or classic logic — and state it openly.
How do you ensure that a prompt still works after a model update?
Prompts are software: without regression tests, quality degrades unnoticed with every provider update.
Show an example where a prompt detail changed the result drastically.
Practitioners have such stories ready immediately — they demonstrate depth beyond standard tips.
What does a Prompt Engineer cost as a freelancer?
Based on our market observations, hourly rates in the DACH region in 2026 range between 85 and 130 euros — profiles with eval methodology and product experience at the upper end.
Won't prompt engineering soon become obsolete?
The role is shifting: hardly anyone still needs to optimise simple everyday prompts. In products, agent systems and specialist processes, however, systematic prompt design with evaluation remains decisive for quality.
Do we need a full-time Prompt Engineer?
Rarely. Compact engagements are typical: building a prompt library, stabilising a product feature or training teams — often 10 to 30 consulting days rather than a permanent position.
What distinguishes a Prompt Engineer from an LLM Engineer?
The Prompt Engineer works on the instructions and their quality measurement; the LLM Engineer builds the application around them — architecture, APIs, operations. In projects, the two roles complement each other.

Looking for a Prompt Engineer?

Tell us about your project — within one business day you’ll get an honest assessment of availability, day rate and alternatives.

Request experts