New opportunity

AI Architect

PACS Prudential Assurance Company Singapore (Pte) Ltd · Singapore

About this role

The AI Architect bridges business strategy and AI technology at Prudential in Singapore. The role works with business leaders, product owners and technology teams to identify and prioritize AI opportunities, translate business challenges into solutions, and deliver measurable outcomes.

Responsibilities

Engage senior leaders to uncover opportunities, advise on AI strategy and adoption, build business cases, and demonstrate value realization.
Envision and design scalable, secure, end-to-end AI, machine learning and Generative AI solutions aligned with enterprise standards. Define reference architectures, reusable capabilities, architecture principles and development methodologies for enterprise-wide adoption.
Lead solution design for Agentic AI, retrieval-augmented generation (RAG), MCP servers, Model APIs and AI orchestration frameworks. Establish practices for deployment, MLOps, LLMOps, monitoring and lifecycle management.

Requirements

5-7+ years of experience in AI, data, solution architecture or enterprise architecture roles; proven experience leading AI discovery, solution design and architecture engagements with business stakeholders, and delivering enterprise AI and Generative AI solutions from concept through implementation.
Strong understanding of machine learning, Generative AI, LLMs, NLP, Agentic AI, governance, security, compliance and responsible AI. Experience with LangChain, LangGraph, RAG frameworks, MCP servers and AI orchestration platforms; knowledge of Azure and GCP AI platforms, MLOps, LLMOps, observability, data architecture, integration patterns and APIs.
Bachelor's or master's degree in Computer Science, Data Science, Engineering, Information Technology or a related discipline.

Advantageous experience includes developing reusable services and AI enablement capabilities, using Databricks and MLflow, and working in insurance, financial services or other highly regulated industries.

For full details about this role, click “Apply on company site” to view the employer’s original posting.

Skills for this role

AI solution architectureMachine LearningGenerative AILarge Language ModelsNatural Language ProcessingAgentic AIRetrieval-Augmented GenerationLangChainLangGraphMCP serversModel APIsAI orchestrationAzureGoogle Cloud PlatformMLOpsLLMOpsModel governanceObservabilityResponsible AIData architectureAPI integrationDatabricksMLflowAI strategySolution designEnterprise architectureAI securityComplianceStakeholder engagement

Checking your profile…

Preparation suggestions, not the employer’s actual interview questions. Always check the original posting for current requirements.

Understand the role

What matters most

Turn business opportunities into measurable AI outcomes
The role engages senior leaders, prioritizes AI opportunities, advises on adoption, builds business cases and demonstrates value realization. For preparation, outline how you would move from discovery to a proposed use case, success measures and adoption approach.
Design secure, scalable end-to-end architectures
The posting calls for AI, ML and Generative AI architectures aligned with enterprise standards, from concept through implementation. Prepare to explain design decisions across data, integration, APIs, cloud platforms and security.
Set reusable enterprise architecture patterns
The architect defines reference designs, reusable capabilities, architecture principles and development methodologies for enterprise-wide adoption. Prepare an example of when a shared pattern is preferable to a use-case-specific design.
Design for deployment and ongoing oversight
The role leads design for Agentic AI, RAG, MCP servers, Model APIs and orchestration, and defines practices for deployment, MLOps, LLMOps, monitoring and lifecycle management. Prepare to connect the solution design to observability, governance and responsible AI.

Likely challenges to think through

Balancing value against implementation effort
A likely challenge is choosing among promising AI opportunities when business value, feasibility and adoption effort differ; this is a preparation hypothesis, not a reported employer problem. Prepare a prioritization approach that makes those trade-offs and outcome measures explicit.
Standardizing without constraining use cases
Enterprise-wide reference patterns may need to accommodate different AI solutions without becoming too rigid. Prepare to explain what you would standardize, what you would leave configurable and how you would evaluate exceptions.
Managing emerging AI within governance boundaries
Agentic AI and RAG can add orchestration and data-access complexity while the posting emphasizes security, compliance and responsible AI. Prepare to discuss design controls, monitoring and lifecycle decisions as architecture trade-offs, without assuming a particular Prudential policy or system.

Practice your answers

Technical practice

How would you prioritize AI opportunities with business leaders?

What to coverThe role includes AI discovery, business cases and measurable outcomes. Suggested answer focus: compare expected value, feasibility, risk and adoption effort, then explain how you would define a baseline and measure results.

When would you choose conventional ML, NLP or Generative AI for a business problem?

What to coverThe posting calls for architectures spanning all three. Suggested answer focus: relate the choice to the task, available data, evaluation criteria and operational risks rather than assuming an LLM is appropriate.

How would you design a scalable, secure end-to-end AI architecture?

What to coverThe role requires enterprise-aligned solution design and knowledge of Azure and GCP AI platforms. Suggested answer focus: describe the major components and security boundaries, then explain how you would weigh platform fit, scalability and enterprise standards.

How would you design and evaluate a RAG solution?

What to coverRAG architecture is explicitly part of the role. Suggested answer focus: cover source-data preparation, retrieval, answer grounding and evaluation of retrieval quality and generated responses, including how you would investigate poor answers.

What would guide your design of an Agentic AI workflow?

What to coverThe posting names Agentic AI, LangChain, LangGraph and orchestration platforms. Suggested answer focus: explain when multi-step orchestration adds value, how you would control tool use and failure paths, and when a simpler workflow would suffice.

How would you integrate MCP servers and Model APIs into an AI solution?

What to coverLeading designs involving MCP servers and Model APIs is a stated duty. Suggested answer focus: discuss interface boundaries, authorization, data exposure, error handling and the trade-off between reusable integration and solution-specific needs.

How would you address governance, security and responsible AI in a proposed design?

What to coverThe posting explicitly calls for these principles alongside compliance. Suggested answer focus: identify risks early, describe proportionate controls and review points, and explain how you would document design decisions without claiming a particular company policy.

What deployment, monitoring and lifecycle practices would you establish for AI systems?

What to coverDefining MLOps, LLMOps, observability and lifecycle practices is part of the role. Suggested answer focus: cover versioning, evaluation before release, monitoring after deployment and a response to deteriorating quality or changing data.

How would you define reusable AI architecture patterns across teams?

What to coverThe role includes reference designs, reusable capabilities and standardization. Suggested answer focus: distinguish what should be standardized from what teams may tailor, and explain how you would assess whether a pattern is secure, usable and adopted.

How would data architecture affect an enterprise AI solution?

What to coverThe posting calls for familiarity with data architecture, integration patterns and scalable enterprise platforms. Suggested answer focus: trace how data is accessed and prepared, address quality and access constraints, and explain the impact of those choices on solution reliability.

Behavioural practice

Describe a time you uncovered an AI opportunity with senior stakeholders.

What to coverThe role involves leading discovery with business leaders. Suggested answer focus: use situation, action and result to show how you clarified the problem, tested its suitability for AI and established a measurable outcome.

Tell me about a time you aligned business, product and technology teams on a solution.

What to coverThe posting emphasizes partnership across these groups. Suggested answer focus: describe the differing priorities, what you did to reach an architecture decision and the resulting implementation or decision.

Give an example of making a trade-off between solution ambition and adoption.

What to coverThe role covers AI strategy, architecture and adoption. Suggested answer focus: explain the situation, the options you presented, how you secured a decision and what happened after it was put into practice.

Describe a time you introduced a reusable architecture pattern or standard.

What to coverEnterprise-wide reuse and design standardization are stated responsibilities. Suggested answer focus: explain the need, how you worked with affected teams and what changed as a result.

Tell me about a time you raised an AI governance or security concern.

What to coverGovernance, security, compliance and responsible AI are explicit areas of the role. Suggested answer focus: set out the situation, your action with relevant stakeholders and the outcome, without implying that any particular control is required by this employer.

Sharpen your knowledge

Topics to prepare

AI opportunity prioritization and value
Review how to compare AI opportunities by business value, feasibility and measurable outcomes. Exercise: outline a hypothetical opportunity, its business case and success measures. The role advises senior leaders on AI strategy, adoption and value realization.
ML, NLP and LLM solution choices
Review when an ML, NLP or Generative AI approach fits a business challenge and what each requires from data and evaluation. Exercise: compare two approaches for a hypothetical use case and explain the trade-offs. The posting calls for strong understanding across these areas.
Enterprise AI architecture patterns
Review scalability, security, reference designs and reusable capabilities. Exercise: critique a hypothetical end-to-end AI architecture against enterprise standards. Defining architecture patterns and principles is an explicit responsibility.
RAG, agents and orchestration concepts
Review how retrieval, Agentic AI, MCP servers, Model APIs and orchestration fit together, including where their boundaries differ. Exercise: explain the component choices for a hypothetical RAG-plus-agent design. The role leads solution design in these areas.
Governance and lifecycle principles
Review responsible AI, security, compliance, model governance, observability, MLOps and LLMOps. Exercise: identify governance and monitoring decisions at each lifecycle stage of a hypothetical solution. These are explicitly named in the posting.
Cloud, data and integration architecture
Review how data architecture, APIs and integration patterns support AI solutions on Azure and GCP. Exercise: compare two high-level deployment designs and explain their integration and scaling trade-offs. The posting requests knowledge of both cloud platforms and familiarity with enterprise integration.

Skills to revise

Lead an AI discovery conversation
Practise a short stakeholder interview: elicit the business problem, constraints, adoption needs and measures of value, then produce a prioritized opportunity brief. This reflects the stated work with senior leaders, product owners and technology teams.
Produce an end-to-end reference design
Draw a hypothetical scalable, secure AI solution from data inputs through Model APIs to deployment and monitoring; annotate reusable components and architecture decisions. The role is responsible for reference designs and enterprise-wide patterns.
Design an agent orchestration flow
Sketch or build a small practice flow with LangChain or LangGraph, showing where an MCP server, Model API and human oversight could fit. Label it as practice rather than a Prudential implementation. The posting explicitly names these technologies and design areas.
Work through a RAG design exercise
For a hypothetical document question-answering task, specify retrieval, response generation and checks for answer quality; explain failure cases. RAG architecture and frameworks are explicit areas of experience.
Draft an operations and governance plan
Write a one-page practice plan covering deployment, monitoring, lifecycle updates, model governance and responsible-AI review for an AI solution. This rehearses the role's responsibility for MLOps, LLMOps and lifecycle practices.

Your preparation checklist

Draft an AI opportunity brief
The role involves senior-stakeholder discovery and prioritization. Prepare a sample brief with the business problem, candidate use cases, assumptions, success measures and a business-case outline.
Prepare an end-to-end project walkthrough
The posting seeks enterprise AI and Generative AI delivery from concept through implementation. If you have a relevant example, document your design decisions, stakeholder engagement, rollout and measured outcomes; distinguish your contribution from the team’s.
Practice a solution-design exercise
Sketch a scalable RAG or agentic AI architecture showing data flow, retrieval, orchestration, Model APIs, integration points and security boundaries. Be ready to explain when LangChain, LangGraph or MCP servers would—and would not—fit.
Create a reusable architecture artifact
The role calls for reference designs and enterprise-wide AI capabilities. Draft a one-page reference architecture or set of design principles, noting assumptions and how it could align with enterprise standards.
Outline operational and governance controls
Prepare a lifecycle diagram covering deployment, MLOps or LLMOps, monitoring, observability, model governance and responsible AI. Identify decisions that would need validation against the employer’s security and compliance requirements.
Organize evidence for the stated requirements
Make a concise evidence matrix for architecture leadership, business-stakeholder work, Azure and GCP knowledge, and relevant tools. Mark Databricks, MLflow and regulated-industry experience as advantageous rather than mandatory; leave any gaps explicit.
Bring targeted questions for the employer
Ask how AI opportunities are prioritized, which enterprise architecture standards apply, what reusable capabilities already exist, and how value realization and AI governance are assessed.

Availability last checked: 4 Nov 2026. Status can change after a check; confirm on the employer’s website. Report an issue · Our checking policy

KEEP EXPLORING

Similar AI jobs

Related skills and specializations in Singapore. Matched to this role, not your profile.

Explore more jobs

Finding similar opportunities…