AI & LLM Application Developer

AI \& LLM Application Developer

We are a progressive, forward-thinking agency group with multiple established businesses and a genuine appetite for innovation.

Our experience spans marketing, media, automotive, technology, content and digital services - but our ambitions extend well beyond any one sector. We are developing AI and LLM-powered products that can transform our businesses, create new commercial services and solve challenges across a wider range of industries.

Our multiskilled team includes specialists from media, marketing, design, content, data, digital, sales and commercial strategy. We enjoy exploring new ideas and turning emerging technology into something practical, reliable and valuable.

We are building an internal AI development team and are looking for an exceptional AI \& LLM Application Developer to join us at the beginning of that journey.

The opportunity

Working closely with our Lead AI \& LLM Developer, you will play a hands-on role in developing, testing and improving new AI-powered products, applications and internal tools.

Your work will focus particularly on advanced prompting, context engineering, guardrails, model evaluation and the creation of reliable LLM workflows.

Although this position reports to a senior technical lead, it is not a basic entry-level AI role. We are looking for someone who already uses Anthropic Claude and other leading platforms exceptionally well and can demonstrate the technical thinking required to turn AI capability into controlled, repeatable processes.

What you will be doing

• Supporting the development of AI and LLM-powered applications
• Working extensively with Anthropic Claude, OpenAI, Google Gemini and other suitable models
• Designing and refining system prompts, context structures and reusable prompt frameworks
• Creating structured outputs and multi-stage AI workflows
• Supporting function calling, tool use and agentic applications
• Building guardrails that control model behaviour
• Testing models against expected, unexpected and deliberately challenging inputs
• Identifying hallucinations, inconsistencies, bias, security risks and failure points
• Testing for prompt injection, data leakage and attempted guardrail bypasses
• Designing fallback responses, escalation routes and human-review processes
• Creating measurable evaluation criteria and repeatable test suites
• Comparing models, prompts and configurations using evidence
• Supporting API integrations and application development
• Assisting with RAG, embeddings, vector databases and knowledge-base preparation
• Testing whether AI answers are properly grounded in approved information
• Monitoring context use, token consumption, response time, reliability and cost
• Maintaining prompt versions, evaluation results and technical documentation
• Supporting multimodal applications involving text, documents, images and audio
• Contributing to voice-AI workflows where relevant
• Working with specialists across our businesses to understand real user requirements
• Supporting controlled user and customer testing
• Investigating faults and improving products following feedback
• Keeping pace with developments across the AI and LLM market

Essential LLM knowledge

You must demonstrate strong practical knowledge of large language models and how they are applied within reliable AI products.

This should include

• Understanding how LLMs process context and generate responses
• Knowledge of tokens, context windows, inference settings and model limitations
• Advanced prompting and context-engineering techniques
• System prompts, structured outputs and tool use
• Comparing Claude, OpenAI, Gemini and other models
• Designing test cases and evaluating output quality
• Identifying hallucinations and inconsistent model behaviour
• Understanding prompt injection and data-leakage risks
• Developing guardrails, fallback behaviour and human-review processes
• Understanding RAG, embeddings and governed knowledge bases
• Awareness of response times, token usage and API costs
• Understanding when an LLM should—and should not—be used

You do not need to have trained an LLM from the ground up. You must, however, understand how to work with LLMs systematically and apply them within controlled applications and business workflows.

Technical capability

You should demonstrate experience or strong practical ability in several of the following:

• Python, JavaScript or another relevant programming language
• Anthropic, OpenAI, Google or comparable APIs
• JSON, API requests and structured data
• Prompt chaining and multi-stage workflows
• Retrieval-augmented generation
• Embeddings and vector databases
• Automated model testing and evaluation
• Model Context Protocol or tool-connected AI systems
• Git and version control
• Data privacy, GDPR and responsible AI
• Clear technical documentation

Experience with LangChain, LlamaIndex, LangGraph, cloud platforms or AI observability tools would be advantageous.

Exceptional AI-platform ability

Basic use of ChatGPT or Claude for content generation will not be sufficient.

Applicants should demonstrate

• Advanced and resourceful use of Anthropic Claude
• Practical experience with other leading models and AI platforms
• The ability to construct complex, controlled prompt workflows
• Understanding of differences between models
• The ability to choose a platform according to quality, risk, speed and cost
• Experience identifying and correcting unreliable outputs
• The ability to learn and apply newly released AI capabilities quickly

The person we are looking for

You will

• Be genuinely excited by AI, innovation and technological change
• Be curious about why models behave in particular ways
• Enjoy experimenting, testing and solving difficult problems
• Identify edge cases and risks that other people may miss
• Test assumptions rather than accepting plausible-looking responses
• Combine creativity with structure and attention to detail
• Have strong written-language and analytical skills
• Communicate clearly with technical and non-technical colleagues
• Take ownership and escalate problems promptly
• Enjoy collaborating with specialists from different disciplines
• Want to grow alongside a new AI team and product portfolio

Show us what you can do

Applicants must provide evidence of genuine practical experience with AI systems.

We would particularly like to see

• A complex LLM workflow, assistant, automation or application you created
• Examples of advanced work using Anthropic Claude
• A difficult prompting or context problem you solved
• How you evaluated and improved model outputs
• Guardrails you designed and the risks they addressed
• An example of identifying and correcting unreliable AI behaviour
• Any Python, API, RAG, structured-output or tool-use work completed
• A project that reached real users or delivered a measurable benefit
• A clear explanation of your individual contribution

Commercial, academic, personal and open-source projects will be considered if they demonstrate exceptional ability and genuine technical understanding.

Why join us?

We offer

• The opportunity to join a new AI development team at an early and influential stage
• Hands-on involvement in products intended for commercial use
• Support and development from a senior Lead AI \& LLM Developer
• Exposure to multiple businesses, sectors and commercial opportunities
• Collaboration with specialists across marketing, media, design, content, digital, data and strategy
• Access to genuine business challenges and potential test users
• The opportunity to shape our prompting, evaluation and AI-safety standards
• Significant progression potential as the team and products grow
• A warm, welcoming and fun working environment
• A collaborative culture where curiosity, ideas and initiative are encouraged
• Office-based collaboration Monday to Thursday
• The option to work from home on Fridays
• Plenty of free on-site parking
• Company benefits and employee perks
• The stability of an established group combined with the energy of a new technology venture

How to apply

Please submit your CV with a portfolio or supporting statement explaining:

• How you have used Anthropic Claude at an advanced level
• Other LLMs and AI platforms you use confidently
• AI applications, assistants or workflows you have created
• Examples of prompts, guardrails and evaluation processes you designed
• The technologies and models used
• Your individual contribution
• Measurable results or practical outcomes
• Your salary expectations and notice period
• Your ability to work from our Hampshire office Monday to Thursday

Shortlisted applicants will be asked to complete a controlled practical assessment involving advanced prompting, model comparison, output evaluation and the design of appropriate guardrails.

Pay: £35,000.00-£55,000.00 per year

Benefits

• Additional leave
• Casual dress
• Company events
• Company pension
• Free parking
• On-site parking
• Sick pay
• Work from home

Work Location: In person