Vacancy Name

AI Agentic Engineer (Contract)

Location(s)

London

Employment Type

Full Time Fixed Term

Department

Carne Technology

Description

Who Are We Looking For?

Carne is seeking an AI Agentic Engineer on a 6-month fixed-term contractor basis to support the design and delivery of production-grade agentic AI solutions across regulated fund operations and wider financial services processes. The role is based in Dublin or London and will report to the Chief Technology Officer.

This is a hands-on engineering role focused on translating business problems into reliable agentic systems that can retrieve information, use tools, coordinate workflows and support decision-making safely and effectively. Typical use cases may include workflow orchestration, decision support, exception handling, process automation, data enrichment and system-to-system coordination across operations, compliance, risk and client servicing.

The contractor will design, build and improve multi-step agentic pipelines that operate in a control-driven environment. This includes selecting the right orchestration approach, integrating models and tools with internal systems and data sources, designing memory and retrieval patterns, and ensuring outputs are accurate, auditable and grounded in source information.

The ideal candidate will bring strong AI engineering judgement rather than demo-level prompt-chaining experience. They should be able to diagnose failures across prompts, retrieval, tool integrations, planning logic and source data; build meaningful evaluation harnesses; instrument production traces; and apply FinOps discipline to manage model usage, latency, token consumption and operational spend.

This role requires someone who can work independently with Product Engineering, Cloud, Cyber Security, business, risk, compliance and operations stakeholders. The successful candidate will design agentic systems that are secure, observable, cost-aware and safe by design, with appropriate governance, auditability, human-in-the-loop controls and escalation paths built in from the outset.

What are the minimum requirements?

• Strong hands-on experience building production-grade AI, software engineering, machine learning, quantitative engineering or closely related systems.
• Recent practical experience designing and delivering agentic AI solutions, ideally including ownership of reliability, evaluation, observability and operational performance.
• Experience designing multi-step agentic pipelines using orchestration approaches and frameworks such as LangGraph, LangChain, LlamaIndex, Semantic Kernel, AutoGen, CrewAI or equivalent custom orchestration layers.
• Strong understanding of agentic design patterns, including orchestrator-worker, planner-executor, supervisor-subagent, router, tool-using agents, event-driven workflows and human-in-the-loop escalation.
• Experience integrating large language models and AI services with internal systems, third-party platforms, document repositories, APIs and structured interfaces.
• Fluency with commercial and/or open-source model providers and deployment approaches, including function calling, structured outputs and enterprise deployment patterns such as Azure OpenAI Service or equivalent.
• Strong Python skills and SQL fluency; TypeScript, C#/.NET or other enterprise integration experience would be advantageous.
• Experience with retrieval and grounding patterns, including enterprise knowledge retrieval, vector search, document-grounded retrieval and structured output validation.
• Experience designing memory architectures spanning short-term task context, retrieval over enterprise knowledge and operational data, and longer-term semantic or state stores.
• Ability to build evaluation harnesses for non-deterministic systems, including task completion, extraction accuracy, tool-call correctness, factual grounding and failure-mode analysis.
• Experience implementing safety guardrails such as output validation, prompt injection defences, confidentiality and PII controls, audit logging and human-in-the-loop escalation triggers.
• Practical FinOps discipline for AI workloads, including token optimisation, latency management, model selection, caching, batching and cost guardrails.
• Experience with observability and traceability tooling such as LangSmith, Weights & Biases, Azure Monitor, Application Insights or equivalent platforms.
• Experience with enterprise cloud and platform services, ideally in a Microsoft-centric environment such as Azure OpenAI Service, Azure AI Foundry, Azure AI Search, Azure AI Document Intelligence, Azure Functions or Logic Apps; equivalent cloud experience is also relevant.
• Understanding of infrastructure patterns required to run production agent systems, including state persistence, API and event integration, containerised deployment, identity, secrets management and CI/CD.
• Experience with DevOps, MLOps or platform engineering practices such as Azure DevOps, GitHub Actions, Terraform, Bicep, Entra ID, Key Vault, model and prompt versioning, and controlled deployment processes.
• Ability to work with business, product, technology, cyber security, risk, compliance, legal and operations stakeholders to translate real-world requirements into agent task design, acceptance criteria and escalation rules.
• Knowledge of financial services, funds, asset management, fintech or another regulated operating environment is advantageous but not essential for candidates with strong engineering depth and sound product judgement.