• Opportunity to join a growing organisation with a strong pipeline of work
• Stability and job security with an established business
• The chance to make an impact across multiple business units
• Supportive, collaborative team environment
• Career progression with clear pathways
• Ongoing training, mentorship and development
• Discounts on technology, cars, fashion and leisure
• Modern systems and tools

Key Duties and Responsibilities:

• Find routine, high-volume work across the business and turn it into AI-assisted workflows that people supervise rather than perform.
• Design, build and maintain custom AI agents, including tool definitions, control flow, state and memory, retries, escalation paths and stopping conditions.
• Integrate LLM and multimodal model APIs into production applications and build structured extraction from PDFs, scans, photos and emails with per-field confidence.
• Write and maintain evaluation sets from real production data, set confidence thresholds, and support staged rollouts backed by evidence.
• Design human-in-the-loop experiences that make AI confidence visible, simplify review and capture feedback for continuous improvement.
• Instrument AI systems with tracing, token usage, cost, latency and quality metrics, and build dashboards that provide meaningful insights.
• Build secure, production-ready AI solutions by managing prompt injection risks, permission boundaries, audit trails and approval gates.
• Build and maintain shared AI capabilities, including MCP servers, tool libraries, prompt and configuration versioning, AI gateways and evaluation tooling.
• Write clear documentation, build maintainable interfaces, and debug and support solutions through production and beyond.

Key Selection Criteria:

• Proven experience shipping LLM-powered systems to real users and supporting them in production.
• Experience designing and building AI agents, including tool design, orchestration, state management and human-in-the-loop workflows.
• Strong Python and TypeScript skills, with experience building production-ready AI applications and services.
• Experience with prompt and context engineering, structured outputs, evaluation frameworks and confidence-based AI systems.
• Solid understanding of Model Context Protocol (MCP), AI security, prompt injection mitigation and responsible AI practices.
• Experience managing model selection, token usage, latency, cost and performance in production environments.
• Strong knowledge of leading AI models, including Google Gemini, Anthropic Claude and OpenAI GPT.
• Hands-on experience with Google Cloud, particularly Vertex AI, Gemini, Cloud Run, BigQuery and related services.
• 5+ years' experience building and shipping production software, including API design, event-driven architecture, CI/CD, infrastructure as code, Git and testing.
• Experience working across existing technology stacks, with the ability to quickly understand and contribute to established codebases.
• Experience with vision and multimodal models, document extraction, AI observability, retrieval, vector search or workflow engines will be highly regarded.

• Opportunity to join a growing organisation with a strong pipeline of work
• Stability and job security with an established business
• The chance to make an impact across multiple business units
• Supportive, collaborative team environment
• Career progression with clear pathways
• Ongoing training, mentorship and development
• Discounts on technology, cars, fashion and leisure
• Modern systems and tools

Key Duties and Responsibilities:

• Find routine, high-volume work across the business and turn it into AI-assisted workflows that people supervise rather than perform.
• Design, build and maintain custom AI agents, including tool definitions, control flow, state and memory, retries, escalation paths and stopping conditions.
• Integrate LLM and multimodal model APIs into production applications and build structured extraction from PDFs, scans, photos and emails with per-field confidence.
• Write and maintain evaluation sets from real production data, set confidence thresholds, and support staged rollouts backed by evidence.
• Design human-in-the-loop experiences that make AI confidence visible, simplify review and capture feedback for continuous improvement.
• Instrument AI systems with tracing, token usage, cost, latency and quality metrics, and build dashboards that provide meaningful insights.
• Build secure, production-ready AI solutions by managing prompt injection risks, permission boundaries, audit trails and approval gates.
• Build and maintain shared AI capabilities, including MCP servers, tool libraries, prompt and configuration versioning, AI gateways and evaluation tooling.
• Write clear documentation, build maintainable interfaces, and debug and support solutions through production and beyond.

Key Selection Criteria:

• Proven experience shipping LLM-powered systems to real users and supporting them in production.
• Experience designing and building AI agents, including tool design, orchestration, state management and human-in-the-loop workflows.
• Strong Python and TypeScript skills, with experience building production-ready AI applications and services.
• Experience with prompt and context engineering, structured outputs, evaluation frameworks and confidence-based AI systems.
• Solid understanding of Model Context Protocol (MCP), AI security, prompt injection mitigation and responsible AI practices.
• Experience managing model selection, token usage, latency, cost and performance in production environments.
• Strong knowledge of leading AI models, including Google Gemini, Anthropic Claude and OpenAI GPT.
• Hands-on experience with Google Cloud, particularly Vertex AI, Gemini, Cloud Run, BigQuery and related services.
• 5+ years' experience building and shipping production software, including API design, event-driven architecture, CI/CD, infrastructure as code, Git and testing.
• Experience working across existing technology stacks, with the ability to quickly understand and contribute to established codebases.
• Experience with vision and multimodal models, document extraction, AI observability, retrieval, vector search or workflow engines will be highly regarded.