Allianz—AI Tech Lead (Data Engineering) · D3006

Job description

Description: Purpose of the Role The AI Lead is a senior technical leadership role responsible for driving the end-to-end delivery of AI and Agentic AI solutions. You will own the technical direction of AI projects, lead a team of AI Engineers, and serve as the bridge between business stakeholders, solution architects, and engineering teams. You are accountable for solution quality, delivery cadence, and production readiness of AI workloads. Key Accountabilities Technical Leadership & Architecture - Own the technical delivery of AI solutions from ideation through to production deployment - Define and enforce solution architecture patterns including single-agent, multi-agent, and orchestration topologies - Make technology selection decisions with cost, performance, and maintainability trade-offs - Design and approve prompt architectures, retrieval strategies, and context assembly patterns - Establish reference architectures and reusable design patterns for the AI engineering team Delivery Management - Lead AI projects through a stage-gated delivery model: Ideation → Framing → Design → Build → Test → UAT → Production - Own design approval gates and ensure every deliverable meets agreed quality bars before stage exit - Manage non-functional requirements (latency, cost-per-task, throughput) and ensure they are met at each gate - Drive parallel workstreams (e.g., systems data readiness alongside agent build) to prevent schedule slip - Coordinate cross-team dependencies with platform, MLOps, data, and business teams Team Leadership - Lead, mentor, and grow a team of AI Engineers across agent development, prompt engineering, and LLM integration - Conduct code reviews and AI engineering quality checks - Establish and enforce development standards, coding guidelines, and engineering best practices - Foster a culture of experimentation, continuous learning, and knowledge sharing - Manage team capacity, sprint planning, and delivery commitments Agentic AI & LLM Expertise - Design and oversee agentic orchestration solutions (multi-agent systems, state management, human-in-the-loop patterns) - Lead prompt engineering strategy including system prompt design, few-shot patterns, chain-of-thought reasoning, and prompt versioning - Architect Retrieval-Augmented Generation (RAG) pipelines including chunking strategies, embedding models, and retrieval quality evaluation - Drive evaluation harness design: golden-set creation, grading rubrics, pass-bar thresholds, and CI-integrated eval suites - Ensure non-determinism is managed through stable eval scores across consecutive runs Security & Governance - Ensure all AI solutions follow zero-trust security principles: managed identities, no secrets in code, defence in depth - Integrate content safety checks on all LLM inputs and outputs - Design and enforce guardrails for agent actions including blast-radius assessment and fallback paths - Own adversarial testing strategy: prompt injection, jailbreak, and tool-misuse test packs - Contribute to governance frameworks including model cards, explainability documentation, and audit evidence Stakeholder Engagement - Translate business requirements into feasible AI solution designs - Present architecture decisions, PoC findings, and delivery progress to senior stakeholders - Collaborate with business SMEs on use-case framing, autonomy-level decisions, and acceptance criteria - Participate in joint go/no-go gates for production deployment Observability & Operations - Define observability and trace design requirements for AI workloads (logging, telemetry, cost tracking) - Establish alert thresholds, on-call procedures, and hypercare plans for production AI systems - Drive token usage monitoring, cost optimisation, and model-tier selection strategies - Own post-deployment monitoring, prompt optimisation, and continuous improvement cycles Required Skills & Experience Essential - 8+ years in software engineering with at least 3 years focused on AI/ML solutions - 2+ years of hands-on experience with Large Language Models (GPT-4 class or equivalent) in production environments - Proven experience designing and delivering agentic AI systems (multi-agent orchestration, tool-calling, state management) - Strong proficiency in Python and modern AI frameworks (LangChain, LangGraph, Semantic Kernel, or equivalent) - Deep 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 cloud AI platforms (Azure AI, AWS Bedrock, GCP Vertex AI, or equivalent) - Experience leading engineering teams (2–6 direct reports) with a track record of delivering production AI systems - Strong understanding of CI/CD for AI workloads including eval-suite integration and automated testing - Knowledge of identity and security patterns: managed identities, RBAC, content safety, zero-trust principles Desirable - Experience with event-driven architectures (Kafka, Service Bus, or equivalent messaging systems) - Familiarity with ML model serving and MLOps (MLflow, model registries, drift detection) - Experience with OCR / Document Intelligence solutions - Knowledge of FinOps practices for AI workloads (token cost tracking, model-tier optimisation, budget governance) - Experience with rules engines (Drools, or equivalent) - Insurance, financial services, or other regulated-industry experience - Familiarity with Infrastructure-as-Code (Terraform, Bicep, or equivalent) Key Competencies Technical Depth Can go deep on LLM internals, agent orchestration, and system design while maintaining a broad architectural view Delivery Focus Drives outcomes through structured delivery with clear gates, quality bars, and accountability Communication Translates complex AI concepts for business audiences; writes clear design documents and architecture decision records Pragmatism Balances innovation with production readiness; knows when to PoC and when to ship Security Mindset Treats security as non-negotiable; designs with defence in depth from day one People Leadership Grows engineers through mentoring, code review, and creating space for experimentation Collaboration Works effectively across business, data, platform, and operations teams with a joint-accountability mindset What Success Looks Like (First 12 Months) - Delivered 2–3 AI/agentic solutions to production with measurable business impact - Established engineering standards, templates, and accelerators adopted by the wider team - Built and maintained an evaluation framework with CI-integrated golden-set testing - Achieved stable production operations with defined SLAs, observability, and incident response - Grown the AI engineering team's capability through mentoring, knowledge sharing, and structured onboarding - Contributed to cross-cutting frameworks (security, observability, governance) used across the organisation 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. Great to have you on board. Let's care for tomorrow. Skills: databricks, spark, data architecture, pyspark, data modelling, scala, delta lake, databricks sql, finops, ai solutions deployment, agentic ai, llm orchestration, langchain, llamaindex, rag patterns, cluster policy management, photon engine optimizations, cost-attribution models, medallion architecture, star schema, snowflake, vault 2.0, etl/elt templates, insurance domain knowledge, domain alignment, databricks unity catalog, data governance, data security, leadership, policy-centric data