Senior Prompt Engineer - German LLM Performance

Welo Global; United States; Remote; USD 120,000 - 170,000; Full time

Job summary

Welo Global seeks a Prompt Engineer to manage the end-to-end migration workflow for transitioning templates to LLM autoraters, using client tools to optimize model performance. Responsibilities include APG/APO tooling, manual prompt drafting for complex templates, and monitoring disagreements with the LLM.

The role focuses on ensuring high-quality, structured outputs and rigorous testing against gold data. The candidate should be a German native speaker with fluent English, hold at least a

Qualifications

• Bachelor’s, Master’s, or Doctorate in Computer Science, Data Science, Computational Linguistics, HCI, Cognitive Science, or related analytical field.
• At least 4 years' experience as Prompt Engineer with LLMs for structured outputs.
• Proficiency in identifying error patterns and analyzing model performance.
• Experience using SQL or other data analytics tools.

Responsibilities

• Utilize APG tools to create baseline prompts for template clusters.
• Run and supervise APO tool; review outputs and flag deadlocks/plateaus.
• Manually draft, test, and refine prompts for complex templates and edge cases.
• Monitor shadowbot runs to ensure disagreements are tracked.
• Evaluate prompt outputs against gold data; compute accuracy metrics (F1, precision, recall).
• Draft Launch Certification Documentation for technical launch readiness.

Skills

German (native)

English (fluent)

Prompt engineering

LLM tuning

Data analysis

SQL

Education

Bachelor’s/Master’s/PhD in CS or related field

Job description

Welo Global seeks a Prompt Engineer to manage the end-to-end migration workflow for transitioning templates to LLM autoraters, using client tools to optimize model performance. Responsibilities include APG/APO tooling, manual prompt drafting for complex templates, and monitoring disagreements with the LLM.

The role focuses on ensuring high-quality, structured outputs and rigorous testing against gold data. The candidate should be a German native speaker with fluent English, hold at least a