Agentic AI Engineer

White Oak; Chester; On-site; GBP 70,000 - 100,000; Full time

Job summary

White Oak is embedding agentic AI across the business, from AI-assisted credit underwriting to NLQ of data.

The Agentic AI Engineer will design, build and run AI agents and reusable skills using SnapLogic, Claude Code and our Azure data platform, ensuring accuracy, explainability and auditable outputs.

Qualifications

• Hands-on experience building LLM-based solutions or AI agents (tool use, retrieval / RAG, multi-step workflows) in a commercial or substantial project setting.
• Strong prompt and context engineering — structured prompts, examples, output schemas and systematic testing.
• Good understanding of semantic layers and how they ground AI — measures, dimensions, relationships and business definitions (e.g. Power BI / Tabular) — and the ability to specify what agents need from them.
• Good SQL and working knowledge of a cloud data platform (Azure preferred).
• Experience with APIs and integration (REST, JSON) and some scripting (Python preferred).
• Awareness of how to evaluate and govern AI output — accuracy, explainability, auditability and data privacy.
• Ability to turn business needs into clear technical requirements and explain AI capability and limitations to non-technical stakeholders.
• Organised and delivery-focused, comfortable progressing several use cases at once.

Responsibilities

• Build, test and deploy AI agents and reusable agent skills (e.g. bank statement analysis, transactional history, sales and campaign automation), primarily in SnapLogic via its native tools/Claude Code.
• Design, version and maintain high-quality prompts, system instructions and context, tested against clear acceptance criteria.
• Work with the wider data team to inform the design of semantic models, the business glossary and metric definitions, so they meet the needs of NLQ and agents.
• Connect agents to data and systems via APIs and MCP, including Power BI semantic models and the Azure data warehouse.
• Translate requirements from underwriters, Credit, Sales, Service and Marketing into scoped, testable agent and skill specifications.
• Evaluate agent quality — build test sets, monitor accuracy, hallucination and drift, and iterate.
• Ensure agent outputs are logged, traceable and auditable, following our AI governance and change control (Dev / UAT / Production) processes.
• Work with third-party delivery partners and IT to move proofs of concept into production.
• Document agents, skills and data dependencies so they can be supported, reused and scaled.

Skills

LLM-based solutions

Agent design & orchestration

Prompt engineering

Semantic models knowledge

SQL

Python

APIs & integration

Governance & compliance awareness

Technical requirements translation

Tools

SnapLogic

Claude Code

Azure data platform

Power BI

Job description

White Oak has completed its Azure migration and is now embedding agentic AI across the business — from AI-assisted credit underwriting to natural-language querying (NLQ) of our data across Credit, Sales and Marketing.

The Agentic AI Engineer will design, build and run AI agents and the context that powers them — prompts, instructions and reusable agent skills — drawing on the semantic models and business definitions built by the wider data team. Working within our AI governance framework and alongside delivery partners, they will take use cases from requirements through to production, primarily using SnapLogic, Claude Code and our Azure data platform, ensuring every agent is accurate, explainable and auditable.

Description of role and key responsibilities

• Build, test and deploy AI agents and reusable agent skills (e.g. bank statement analysis, transactional history, sales and campaign automation), primarily in SnapLogic via its native tools/Claude Code.
• Design, version and maintain high-quality prompts, system instructions and context, tested against clear acceptance criteria.
• Work with the wider data team to inform the design of semantic models, the business glossary and metric definitions, so they meet the needs of NLQ and agents.
• Connect agents to data and systems via APIs and MCP, including Power BI semantic models and the Azure data warehouse.
• Translate requirements from underwriters, Credit, Sales, Service and Marketing into scoped, testable agent and skill specifications.
• Evaluate agent quality — build test sets, monitor accuracy, hallucination and drift, and iterate.
• Ensure agent outputs are logged, traceable and auditable, following our AI governance and change control (Dev / UAT / Production) processes.
• Work with third-party delivery partners and IT to move proofs of concept into production.
• Document agents, skills and data dependencies so they can be supported, reused and scaled.

Core skills and knowledge

• Hands-on experience building LLM-based solutions or AI agents (tool use, retrieval / RAG, multi-step workflows) in a commercial or substantial project setting.
• Strong prompt and context engineering — structured prompts, examples, output schemas and systematic testing.
• Good understanding of semantic layers and how they ground AI — measures, dimensions, relationships and business definitions (e.g. Power BI / Tabular) — and the ability to specify what agents need from them.
• Good SQL and working knowledge of a cloud data platform (Azure preferred).
• Experience with APIs and integration (REST, JSON) and some scripting (Python preferred).
• Awareness of how to evaluate and govern AI output — accuracy, explainability, auditability and data privacy.
• Ability to turn business needs into clear technical requirements and explain AI capability and limitations to non-technical stakeholders.
• Organised and delivery-focused, comfortable progressing several use cases at once.

Any other attributes that would be helpful, but not essential for the role

• SnapLogic experience, particularly AgentCreator or agentic pipelines.
• Experience with Model Context Protocol (MCP) or agent frameworks (e.g. Claude, Azure AI Foundry, LangGraph, Semantic Kernel).
• Financial services or lending background — credit underwriting, bank statement or bureau data, FCA-regulated environments.
• Knowledge of AI governance standards and regulation (e.g. EU AI Act, ISO/IEC 42001).
• Exposure to CRM and marketing data (Salesforce, HubSpot) and campaign use cases.
• Curious and pragmatic, with a drive to keep pace with fast-moving AI tooling.