Senior AI Automation Engineer – Digital

The Senior AI Engineer (Agents & Automation) designs, builds, evaluates and operates the AI agents, tools and pipelines that put AI to work across Forever New with e-commerce and digital as the primary focus, and the whole business as the canvas. 

This is a hands-on engineering role at the heart of our AI Centre of Excellence. You will ship production-grade AI agents and, just as importantly, build the enterprise-grade rollout machinery around them: a scalable, evaluated, versioned and governed way of releasing agents, prompts, tools, MCP connectors, RAG and data pipelines. You care as much about how reliably and repeatably we ship and operate AI as about the agents themselves. 

You will operate in an agent-led delivery model where AI agents are trusted to do meaningful work under human supervision with the emphasis on outcomes, quality controls and repeatability rather than approving every step. You'll reengineer the process before you automate it, document what you build, and feed reusable patterns and guardrails back into Forever New's enterprise AI standards. 

In this newly created role you will help move Forever New from promising pilots to a governed, compounding portfolio of agents that genuinely add value no fluff, no hype, no rework. 

• Design and ship production AI agents across the taxonomy conversational/customer-facing, knowledge & Q&A, workflow/process automation, decision-support and monitoring  

• Reengineer and optimise the underlying business process first (mapping it in BPMN), then automate it because automating a broken process just makes waste faster. 

• Integrate tools and MCP connectors with least-privilege, allow-listed access and clean separation between agents that read and agents that act. 

• Build retrieval-augmented generation (RAG), vector and data pipelines grounded in governed data, so agents answer from a single source of truth. 

• Contribute to creating agents that generate dashboards, answer business questions and release insights from enterprise data and the governed semantic layer on Databricks, with evaluation, cost monitoring and release controls built into the delivery process. 

• Data science and machine learning experience is a strong plus, particularly where it supports forecasting, segmentation, recommendation, decision-support or customer/retail intelligence use cases. 

Make the rollout enterprise-grade 

• Build the evaluation harness: quality, groundedness, safety, bias and prompt-injection testing, combining LLM-as-judge with human review, so agents earn their release. 

• Establish versioning and CI/CD for everything agents, prompts, tools, MCPs and pipelines with environments, release gates, rollback and kill-switches. 

• Build with security, privacy and governance by design human-in-the-loop by default for anything touching customers, money or systems of record; full logging and traceability. 

• Manage performance, reliability and cost in production; reduce recurring issues through standard patterns. 

• Document thoroughly runbooks, agent cards, decision records so what you build is supportable and reusable. 

• Partner with and coach business AI champions; where a high-value area is stuck, embed temporarily to deliver the first agent and hand it back. 

• Contribute SME insight, reusable patterns and delivery learnings into Forever New's enterprise AI standards and guardrails through established architecture and governance forums.