Prompt Engineer – Policy.AI (Government Efficiency Accelerator Programme)
Location: London (hybrid work pattern)
Contract type: 3 Months
Rate: £650-£750 inside IR35
This programme is a flagship cross-government initiative driving operational efficiencies through the safe and responsible adoption of AI.
The mission is to simplify workflows, automate routine tasks, and improve productivity across government – delivering better public services and substantial cost savings. The programme will create scalable playbooks and exemplars that enable rapid replication of success across departments.
The Role:
- As a Prompt Engineer, you’ll sit at the forefront of AI innovation in government. You’ll design, test, and optimise prompts for large language models (LLMs) powering copilots and autonomous agents, ensuring they deliver accurate, safe, and contextually relevant outputs.
- You will work alongside product, engineering, and research colleagues to develop intelligent, scalable tools that unlock new levels of productivity for public servants.
Key Responsibilities:
- Design, implement, and optimise AI agents using Copilot, planning algorithms, and decision-making frameworks.
- Develop agent architectures that support autonomy, interactivity, and reliable task completion.
- Integrate agents into real-world workflows, applications, and APIs (e.g. chatbots, copilots, automation tools).
- Collaborate with cross-functional teams to iterate and improve agent performance.
- Monitor and evaluate model behaviour, building safety mechanisms and feedback loops.
- Analyse conversational logs and model outputs to identify opportunities for improvement.
- Maintain robust documentation of design decisions, dependencies, and prompt libraries.
- Champion responsible AI practices, ensuring outputs are safe, ethical, and aligned with public service values.
Skills / Experience required:
You are a creative technologist who thrives at the intersection of AI, design, and systems thinking. You enjoy solving complex problems, collaborating across disciplines, and building tools that make a tangible difference.
You’ll bring experience in:
- Designing and refining prompts for LLMs (e.g. OpenAI, Anthropic, Mistral) to improve accuracy, reliability, and user experience.
- Building and maintaining prompt libraries, templates, and documentation.
- Using data analysis and evaluation frameworks to measure prompt performance.
- Collaborating with engineers, researchers, and product teams on AI-driven initiatives.
- Staying current with emerging techniques in prompt engineering, agentic AI, and LLM optimisation.
- Communicating complex findings to technical and non-technical audiences.
Desirable (but not essential):
- Experience with autonomous agent frameworks (e.g. AutoGPT, CrewAI, LangGraph, ReAct, BabyAGI).
- Background in human–AI interaction, conversational UX, or simulation environments.
- Familiarity with vector databases and retrieval-augmented generation (RAG).
- Understanding of the ethical, social, and safety implications of AI deployment in government.
Please apply online with your CV.
