Agentic AI Engineer - Rapid Prototyping

AI Innovation Engineer £500 per day - Inside IR35 6-12 month contract | London / Fully Remote PCR Digital is working with a large global technology business building a small, highly autonomous AI & Automation Innovation Lab within its commercial organisation.

We're looking for a hands-on AI Innovation Engineer to rapidly build, test and improve AI and automation solutions against real business problems. This is a build-first role. You'll work directly with Commercial and Operations teams, turning ideas into working prototypes and getting them in front of users quickly.

You'll be expected to experiment, build and iterate yourself rather than write requirements for someone else to deliver.

What you'll be doing Build AI agents, automations and LLM-powered prototypes against real business use cases.<br />Prototype using AWS Bedrock and other emerging AI platforms and frameworks.<br />Design and refine prompts, agent instructions, RAG approaches and evaluation criteria.<br />Query and prepare data, work with APIs and connect prototypes to enterprise systems.<br />Run pilots with real users, measure results and iterate quickly based on evidence and feedback.<br />Explore emerging agent frameworks, AI tooling and automation platforms outside the existing technology stack.<br />Compare new technologies against existing platforms across capability, cost and integration effort.<br />Document successful builds and help transition viable pilots towards wider delivery.

What we're looking for The most important thing is that you've personally built AI-enabled solutions, agents or automations. This isn't an AI strategy or programme management role. You'll ideally have strong experience across a combination of:

Generative AI, LLMs and Agentic AI<br />Building AI agents / agentic workflows<br />AWS Bedrock, Azure OpenAI or comparable enterprise LLM platforms<br />Agent frameworks such as LangGraph, LangChain, Semantic Kernel, AutoGen, CrewAI or similar<br />RAG, tool/function calling and grounding LLMs against enterprise data<br />Prompt and agent instruction design<br />LLM evaluation, guardrails, observability and security<br />Strong Python and SQL<br />REST APIs, JSON, webhooks and enterprise integrations<br />Querying and manipulating data for AI use cases<br />Rapid prototyping and taking ideas from POC through to real-user testingThe existing enterprise environment includes AWS/Bedrock, Microsoft, Salesforce and Snowflake.

You don't need expertise across every platform, but you should be comfortable integrating with unfamiliar systems and picking up new technology quickly. The environment You'll be one half of a two-person Innovation Lab, working alongside an AI Innovation Manager. It's deliberately small, autonomous and designed to move quickly.

A major part of the role is exploring emerging AI technology outside the existing stack - from new agent frameworks and foundation models to AI coding, automation, evaluation and observability tools. We're looking for someone who enjoys experimenting with new technology but can apply it to real enterprise problems.

Build quickly, test with users, measure whether it works and move successful ideas towards productionisation