IT Product Owner, BAIC

DUTIES & RESPONSIBILITES

1. Product Vision, Strategy & Roadmap

• Define and own the vision, strategy and roadmap for the agent platform, aligned with the BAIC platform strategy and AstraZeneca’s enterprise AI strategy

• Track the fast-moving agent ecosystem (models, protocols, frameworks, open standards) and decide what to build, buy, adopt or retire; keep the platform modular so components can be swapped as technology changes

• Translate the needs of agent developers and priority business use cases into an outcome-based roadmap with clear OKRs, release plans and success metrics

• Communicate the roadmap and trade-offs clearly to BAIC leadership, developer communities and stakeholders

2. Core Agent Capabilities: LLM Gateway, MCP & Memory

• LLM gateway: own a single, governed entry point to multiple models, with model catalog, routing and fallback, authentication, quotas and rate limits, caching, usage metering and chargeback, logging and content safety controls

• MCP & tools: own the MCP server registry and gateway, standards for building and certifying MCP servers, and secure identity-aware connectors to AZ systems and data (e.g. R&D data platforms, document and knowledge repositories, business applications)

• Memory & context: own short- and long-term agent memory, retrieval (RAG) and vector services, and context management, with clear controls for data access, retention and privacy

• Shape adjacent capabilities as the platform matures: agent runtime and orchestration, agent identity and permissions, agent-to-agent communication, prompt and configuration management

3. Developer Experience & Adoption

• Treat agent developers as customers: run discovery and feedback loops, and measure developer satisfaction and time-to-first-agent

• Deliver a strong self-service developer experience: SDKs and APIs, reference architectures, templates, documentation, sandbox environments and simple onboarding

• Build an internal developer community through demos, office hours, training and champions; drive reuse of shared MCP servers, memory and patterns across teams

• Track platform adoption (teams, agents in production, API and token usage) and use the data to prioritise the backlog

4. Trust, Security & Governance by Design

• Embed guardrails in the platform: prompt-injection and data-leakage protection, PII handling, content filtering, tool permission scoping and human-in-the-loop controls

• Provide observability and evaluation: end-to-end tracing of agent and tool calls, audit trails, quality and safety evaluations, and production monitoring

• Make Responsible AI and compliance easy to follow: align with AZ AI governance, information security and privacy policies, support CSV/GxP needs for agents used in regulated processes, and meet China requirements on generative AI services, data security and cross-border data transfer

• Work with security and architecture teams on threat modelling, access reviews and risk acceptance for new capabilities

5. Backlog Ownership & Roadmap Execution

• Own and prioritise the product backlog; write epics, user stories and acceptance criteria, including non-functional requirements (latency, availability, scalability, security)

• Lead Agile delivery with platform engineering squads and vendors/OSPs; make timely scope and release decisions and deliver the roadmap predictably

• Define and manage platform SLAs/SLOs, versioning and deprecation policies, and backward-compatibility commitments to developers

• Manage dependencies with cloud, identity, network, data and security teams; manage risks, issues and release readiness

6. Value, Cost & Vendor Management

• Own the platform budget (annual plan, MTP inputs, forecasting, variance analysis), including model and token spend

• Apply FinOps for AI: monitor token and compute consumption, optimise model routing and caching, and give consuming teams cost transparency and showback/chargeback

• Track and report platform value (e.g. development time saved, reuse, cost per agent, reliability) to BAIC leadership

• Manage model provider and vendor relationships, including commercial evaluation, contracts, SLAs and performance reviews

JOB REQUIREMENTS & COMPETENCIES

Job requirements

Academic / Professional qualification

• Bachelor’s degree in Computer Science, Software Engineering, Artificial Intelligence, Data Science or a related field (essential)

• Master’s degree in Computer Science, Artificial Intelligence or a related field (desirable)

Technical / skill training

• Product ownership certification (e.g. CSPO, PSPO, SAFe POPM) preferred

• Cloud certification (e.g. AWS or Azure) and recent training in LLMs, agentic AI or AI security is a plus

Working experience

• 6+ years in software engineering, platform engineering or technical product roles, including 2+ years as product owner or product manager for a developer-facing platform, API or cloud service, and hands-on experience with LLM and agent technologies

Essential

• Proven experience owning the roadmap and backlog of a technical platform used by developers, and delivering it with Agile methods

• Hands-on understanding of LLM and agent architecture: LLM gateways and model routing, tool calling and MCP, RAG and vector databases, agent memory, and agent frameworks

• Solid knowledge of cloud-native platforms (e.g. AWS/Azure, containers, Kubernetes), API design, identity and access management (OAuth/OIDC) and observability

• Understanding of AI security and risk (prompt injection, data leakage, tool misuse) and how to mitigate them at platform level

• Experience defining SLAs/SLOs, usage metrics and cost models for shared services

• Able to engage credibly with engineers, and translate technical capabilities into business value for leadership

• High autonomy and ownership; comfortable making decisions in a fast-changing technology landscape

Desirable

• Experience in pharma, life sciences or other regulated industries, and awareness of how agents can support R&D use cases

• Has personally built and deployed agents or MCP servers; contributions to open-source agent or developer tooling

• Experience with LLM evaluation, LLMOps and agent observability tools

• Familiarity with the model ecosystem available in China and with China regulations on generative AI services and data security

• Experience with FinOps for AI workloads and with commercial negotiations with model providers

• Experience in global, matrixed organisations

Language ability

• Fluent English and Mandarin Chinese, written and spoken; able to present to senior international audiences

Computer literacy

• Able to read and write code (e.g. Python or TypeScript) to prototype, review APIs and test platform capabilities

• Proficient with Agile and product tools (e.g. Jira, Confluence), developer tools (e.g. Git, API clients) and everyday GenAI and coding assistants

Others

Key competencies:

• Platform and developer-experience product mindset

• Technical depth and systems thinking

• Clear prioritisation and decision-making under uncertainty

• Execution focus and delivery discipline

• Stakeholder management and influencing without authority

• Curiosity and continuous learning in AI

• Lives AstraZeneca’s values

Date Posted

Closing Date

AstraZeneca embraces diversity and equality of opportunity. We are committed to building an inclusive and diverse team representing all backgrounds, with as wide a range of perspectives as possible, and harnessing industry-leading skills. We believe that the more inclusive we are, the better our work will be. We welcome and consider applications to join our team from all qualified candidates, regardless of their characteristics. We comply with all applicable laws and regulations on non-discrimination in employment (and recruitment), as well as work authorization and employment eligibility verification requirements.