> Markdown version of [/jobs/ext/4679510-ai-agent-engineer](https://www.wearedevelopers.com/jobs/ext/4679510-ai-agent-engineer). Every page supports `.md` or `Accept: text/markdown`. Links point to the HTML versions so they work for humans too. Agent guide: [/agents.md](https://www.wearedevelopers.com/agents.md).

---

# AI Agent Engineer

- **Company:** Digital Waffle
- **Location:** Birmingham, United Kingdom
- **Experience:** Experienced
- **Salary:** £49,993
- **Contract:** Permanent contract
- **Skills:** PHP (Programming Language), Application Programming Interfaces (APIs), Cursor, Commercial Software, TypeScript, Retrieval-Augmented Generation, Large Language Models, Claude Code, Prompt Engineering, Virtual Agents, Api Management
- **Published:** November 12, 2026
- **Apply:** https://www.adzuna.co.uk/jobs/details/5962347257

## About the Role

* 3+ years commercial software engineering experience with strong TypeScript skills (and ideally a background in PHP)
 * Experience with either Claude Code, Cursor or Codex
 * Built and shipped at least one LLM-powered feature or agent into production and operated it afterwards (real, used side projects count)
 * Practical grasp of the agent toolkit: function and tool calling, RAG, vector databases, structured output, prompt engineering and context management
 * Comfortable with at least one agent framework or SDK, with an understanding of why the abstractions exist rather than blind loyalty to a specific one
 * Solid API and integration experience, since much of this role is plumbing agents into systems that were not designed for them
 * Able to talk to non-technical colleagues about what an agent can and cannot reliably do, and confident enough to push back when a request is a bad fit

This role would suit an engineer who wants genuine greenfield ownership of an agent platform, a varied problem space spanning editorial, rights, archive and commercial, and short feedback loops with the people whose work the agents actually change.

## Description

* Design, build and ship production LLM agents against real editorial, rights, archive and commercial workflows, from scoping through to deployment and monitoring
 * Build the shared platform underneath the agents including orchestration, tool and API integrations, retrieval, prompt management, evaluation and observability
 * Integrate agents with existing systems including the archive, submission portal, clips platform and internal editorial tooling
 * Work directly with editorial, research and commercial teams to identify the workflows worth automating and define what 'good enough to trust' looks like for each
 * Build evaluation and guardrails so agent output quality is measurable rather than anecdotal, with human review at the right points
 * Own cost and latency, keeping model spend proportionate to the work being replaced
 * Document and hand over so the platform is not a single-person dependency
 * Keep an eye on the tooling landscape and make pragmatic build-versus-buy calls