Prompt Engineer (44445)

Shaping how AI agents behave in a conversational banking app is the core of this role. As a Prompt Engineer on the AI platform of an international banking group, you will design system prompts and agent instructions, build evaluation datasets and measurably improve LLM applications across markets. Hands-on experience developing prompts for LLM applications, configuring agents and evaluating their behavior is required, together with Python and professional English.

🚀 Project
- design and maintenance of system prompts, agent instructions, routing and handover instructions
- reusable prompt components and rules for knowledge and tool use
- representative evaluation datasets, scoring criteria and regression cases
- failure analysis based on conversation traces and evaluation results
- testing how prompt or model changes affect response quality
- factual grounding, appropriate tool use, language consistency and handling of ambiguous requests
- collaboration with platform engineers, evaluation specialists, product and banking domain teams

🎯 Skills
- hands-on prompt development for LLM applications, incl. production use
- configuring AI agents and evaluating behavior beyond individual examples
- system instructions, context management, few-shot examples, tool calling, retrieval, structured outputs and JSON schemas
- inspecting API requests, tool results and conversation traces
- reproducible experiments and clear documentation of changes
- Python
- English C1

💡 Nice to have
- Git and automated evaluation frameworks
- LLM judges
- LangChain, LangGraph or comparable agent frameworks
- regulated or enterprise environments
- testing multilingual applications

#LI-MV1