AI Engineer - Agentic AI & LLM Solutions

Role : AI Engineer (Agentic AI & LLM Solutions).

Team : CDC AI Transformation.

Focus : Cloud and Data Center.

Purpose of the Role :

The AI Engineer designs, builds, tests, and operates AI and Agentic AI solutions in support of the CDC AI Transformation team.

CDC is focused on cloud and data center services, so the role applies LLMs and agents to the planning, operation, and optimisation of that environment.

You will work as a senior individual contributor inside a delivery team.

You take ownership of your own solutions and workstreams, from design through to production, and work closely with solution architects, platform teams, data teams, and business stakeholders.

You are accountable for the quality, security, and production readiness of what you build.

The role does not carry people-management or team-leadership responsibility.

Key Responsibilities :

Agentic AI & LLM Solution Development :

- Design and build agentic solutions, including single-agent, multi-agent, and orchestration topologies.

- Implement tool-calling, state management, memory, and human-in-the-loop patterns.

- Integrate LLMs into business and operational workflows through APIs, services, and event-driven components.

- Build solutions for cloud and data center use cases, such as incident triage, capacity planning support, infrastructure knowledge assistants, ticket automation, and operational insights.

- Select models and model tiers based on cost, latency, and quality trade-offs.

- Prototype quickly, run PoCs, and take proven concepts through to production.

Prompt Engineering & Context Design :

- Design system prompts, few-shot patterns, and chain-of-thought approaches.

- Build context assembly and retrieval strategies for each use case.

- Version, test, and maintain prompts as managed engineering artefacts.

- Optimise prompts continuously based on production feedback and evaluation results.

RAG & Data Integration :

- Build Retrieval-Augmented Generation (RAG) pipelines : ingestion, chunking, embedding, indexing, and retrieval.

- Select and configure vector databases and embedding models.

- Measure and improve retrieval quality using defined metrics.

- Work with data teams to prepare and connect sources such as runbooks, CMDB data, monitoring data, tickets, and technical documentation.

Evaluation & Testing :

- Build evaluation harnesses : golden sets, grading rubrics, and pass-bar thresholds.

- Integrate eval suites into CI/CD pipelines so quality is checked automatically.

- Manage non-determinism by checking that eval scores are stable across consecutive runs.

- Write unit, integration, and end-to-end tests for agent behaviour and tool use.

- Support UAT with business users and fix issues found.

Security, Safety & Governance :

- Apply zero-trust principles : managed identities, no secrets in code, defence in depth.

- Integrate content safety checks on all LLM inputs and outputs.

- Implement guardrails for agent actions, including blast-radius limits and fallback paths, particularly where agents interact with infrastructure.

- Build and run adversarial test packs : prompt injection, jailbreak, and tool misuse.

- Produce the documentation governance needs, such as model cards, explainability notes, and audit evidence.

- Follow applicable regulation and internal policy (for example EU AI Act, GDPR, DORA) and flag registration or inventory obligations to the right owners.

Deployment, Observability & Operations :

- Build and maintain CI/CD pipelines for AI workloads.

- Implement logging, tracing, telemetry, and cost tracking for AI solutions.

- Set up alerts and support hypercare and on-call arrangements for production systems.

- Monitor token usage and apply cost optimisation and model-tier selection.

- Handle post-deployment monitoring, incident analysis, and continuous improvement.

Collaboration & Engineering Practice :

- Work with solution architects on design decisions and write clear design documents and ADRs.

- Translate business requirements from CDC stakeholders into working technical solutions.

- Give clear progress updates, estimates, and risk flags on your own workstreams.

- Take part in design reviews, code reviews, and joint go/no-go gates.

- Follow and help improve team coding standards and engineering best practices.

- Share knowledge, reusable components, and lessons learned with other engineers.

Required Skills & Experience :

- 5+ years in software engineering, with at least 2 years focused on AI/ML solutions.

- Experience building agentic AI systems (multi-agent orchestration, tool-calling, state management).

- Strong proficiency in Python and modern AI frameworks (LangChain, LangGraph, Semantic Kernel, or equivalent).

- Solid understanding of prompt engineering : system prompts, context assembly, retrieval strategies, and evaluation methods.

- Experience with RAG architectures : vector databases, embedding models, chunking strategies, retrieval quality metrics.

- Hands-on experience with a cloud AI platform (Azure AI, AWS Bedrock, or equivalent).

- Working knowledge of CI/CD for AI workloads.

- Knowledge of identity and security patterns : managed identities, RBAC, content safety, zero-trust.

- Ability to own a workstream, manage your own deliverables, and work independently with limited supervision.

Desirable Skills :

- Cloud and data center experience : strong understanding of cloud infrastructure (compute, networking, storage, hybrid and on-premises environments).

- Experience applying AI to infrastructure, IT operations, or AIOps use cases (monitoring, incident management, capacity forecasting, automation).

- Familiarity with ITSM tooling and processes (ServiceNow or equivalent).

- Experience with event-driven architectures.

- Knowledge of FinOps practices for AI workloads (token cost tracking, model-tier optimisation, budget governance).

- Familiarity with Infrastructure-as-Code (Terraform, Bicep, or equivalent).

- Cloud certifications (Azure, AWS) is +.

Key Competencies :

- Technical Depth : Goes deep on LLM behaviour, agent orchestration, and system design.

- Ownership : Takes responsibility for solution quality and production readiness from design to operation.

- Infrastructure Awareness : Understands cloud, data center, and storage constraints and how they shape AI solution design.

- Pragmatism : Balances innovation with production readiness and knows when to PoC and when to ship.

- Security Mindset : Treats security as non-negotiable and designs with defence in depth from day one.

- Communication : Explains technical concepts clearly to business audiences and writes clear design documents.

- Collaboration : Works across business, data, platform, and operations teams with a joint-accountability mindset.

- Continuous Learning : Keeps up with fast-moving AI tooling and shares what they learn.

Allianz Group is one of the most trusted insurance and asset management companies in the world.

Caring for our employees, their ambitions, dreams and challenges, is what makes us a unique employer.

Together we can build an environment where everyone feels empowered and has the confidence to explore, to grow and to shape a better future for our customers and the world around us.

At Allianz, we stand for unity : we believe that a united world is a more prosperous world, and we are dedicated to consistently advocating for equal opportunities for all.

And the foundation for this is our inclusive workplace, where people and performance both matter, and nurtures a culture grounded in integrity, fairness, inclusion and trust.

We therefore welcome applications regardless of ethnicity or cultural background, age, gender, nationality, religion, social class, disability or sexual orientation, or any other characteristics protected under applicable local laws and regulations.

Skills

Artificial Intelligence, AI Integration, LLM, Agentic AI, Python, LangChain, LangGraph, RAG, Cloud