Document manual workflows and translate them into clear, structured system prompts for AI agents.
Design, iterate, and maintain prompts to improve accuracy, consistency, and guardrails.
Analyse and debug incorrect, inconsistent, or unsafe LLM outputs, identifying whether issues come from prompt design, tools, data or queries, or model limitations.
Collaborate with engineers to understand AI agent tools, backend behaviour, and request improvements.
Work closely with stakeholders to validate workflows and clearly explain agent logic and execution steps.
Test and compare AI agents across multiple LLM providers and configurations.
Maintain prompts, configurations, and documentation in Git-based repositories using Markdown.
Requirements
Mid to senior experience as a Software Engineer or Data Engineer, with strong understanding of backend systems and data workflows.
Strong SQL skills, including writing and debugging queries to ensure correct data outputs.
Ability to read and make small fixes in Python code to support code execution and debugging.
Comfortable reading API documentation and working with structured JSON inputs and outputs.
Proficient in Markdown for writing system prompts, configurations, and documentation.
Able to translate real-world business processes into clear, step-by-step logic that AI agents can reliably follow.
Strong problem-solving skills, including checking logs and documentation to diagnose and fix agent failures.
Comfortable working independently in a small team with end-to-end ownership.
Experience with multiple LLM providers (e.g. OpenAI, Google Gemini, Anthropic Claude) is a plus.
Experience with analytical databases such as Google BigQuery and ability to read backend code (Go is a plus).
Clear communicator, comfortable explaining agent logic, prompt structure, and execution flow to both technical and non-technical stakeholders, and presenting demos when required.
