New opportunity

AI Architect

Fugro · Singapore

About this role

Fugro seeks an AI Architect to design and maintain the architecture for its AI-enabled Marine Ground Digital Twin Platform and guide secure, scalable, cost-efficient AI deployments across APAC. This is an on-site, three-year fixed-term contract in Singapore, open only to Singapore citizens and permanent residents.

Responsibilities

Define technical standards, design patterns and guardrails across models, data pipelines, APIs, cloud infrastructure and security. Evaluate build-versus-buy-versus-integrate options and produce solution designs, data flows, integration maps and technical roadmaps.
Architect Azure, AWS and other compliant cloud solutions, including multi-cloud or hybrid environments where regulations require them. Address APAC data sovereignty, cloud costs, reliability, reproducibility, access control and operational support.
Govern AI and machine-learning environments for experimentation, deployment and monitoring. Assess foundation models, fine-tuning and RAG approaches; design MLOps and LLMOps pipelines covering versioning, evaluation, monitoring and retraining triggers. Set a regional platform strategy using shared services and reusable components.
Embed encryption, access controls and audit logging; align with global governance, work with security and risk stakeholders, and assess and mitigate technical risks. Provide technical direction to the delivery team.

Requirements

A bachelor’s or master’s degree in IT, computer science, engineering or data/AI, or equivalent experience; at least 8 years of experience in solution, software, data or enterprise architecture; and 2+ years providing technical leadership in collaborative or agile delivery. Candidates need experience with governed datasets and data platforms, lakes, lakehouses or warehouses; SQL and Python; large-dataset analysis and preparation; data quality and governance; and security-by-design, privacy-by-design and access control. The role also calls for stakeholder collaboration, experience sponsoring AI and IT initiatives and automation at scale, business acumen, financial and risk governance, and experience in multicultural APAC environments.

Preferred

Familiarity with RAG, model integration, agentic workflows, Microsoft Fabric, Databricks, Azure data services, MLOps or LLMOps, API design, DevOps, infrastructure-as-code, vector databases, semantic search, fine-tuning, generative AI evaluation, Power Platform, Copilot Studio or Azure AI Foundry. Relevant Azure, AI, Databricks or AWS certifications are also preferred. Apply through Fugro’s Careers website with a current resume.

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

Skills for this role

AI architectureSolution architectureData architectureCloud architectureAzureAWSMulti-cloud architectureHybrid cloudData sovereigntyData governanceData lakesLakehousesData warehousesSQLPythonMachine learningFoundation modelsFine-tuningRAGMLOpsLLMOpsModel evaluationModel monitoringSecurity-by-designPrivacy-by-designAccess controlEncryptionAudit loggingRisk assessmentMicrosoft Fabric tooling with Databricks and Azure data services including Power

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

Own the digital twin architecture
Design and maintain the AI-enabled Marine Ground Digital Twin Platform architecture. Set standards across models, data pipelines, APIs and infrastructure; guide delivery teams with solution designs, integration maps and roadmaps.
Design APAC-ready cloud deployments
Architect secure, reliable and cost-efficient deployments across Azure, AWS and other compliant options. Account for country-specific data sovereignty requirements and use multi-cloud or hybrid designs where regulations demand them.
Establish governed data foundations
Support AI use cases with governed datasets from data platforms, lakes, lakehouses or warehouses. The role calls for large-dataset preparation, data quality and governance, with SQL and Python used for access and analysis.
Set the regional AI platform strategy
Define shared services and reusable components; evaluate foundation models, fine-tuning and RAG by use case. Design MLOps and LLMOps for versioning, evaluation, monitoring and retraining triggers.
Embed security and risk controls
Build encryption, access controls and audit logging into architectural decisions. Align with global governance, work with security and risk stakeholders, and define mitigations for new AI initiatives.

Likely challenges to think through

Regional consistency versus local compliance
Preparation hypothesis: Shared APAC architecture may need different deployment patterns when sovereignty rules vary by country, increasing cost and operational complexity. Prepare a decision framework comparing compliant placement, reliability, supportability and spend.
Reusable platform versus use-case fit
Preparation hypothesis: Common AI services can improve consistency, but a particular use case may warrant a different model, RAG design or fine-tuning approach. Prepare to explain evaluation criteria and how versioning, monitoring and retraining decisions would remain governed.
Architectural control versus delivery speed
Preparation hypothesis: Enforcing standards and security guardrails while teams build the digital twin platform could slow delivery if decisions are too rigid. Prepare examples of proportionate guardrails, build-versus-buy-versus-integrate analysis and architecture artefacts that help teams make decisions.

Practice your answers

Technical practice

How would you structure the architecture for Fugro’s AI-enabled Marine Ground Digital Twin Platform?

What to coverPreparation suggestion: Walk through the boundaries between governed data, AI components, APIs, cloud infrastructure and security controls without assuming particular data sources. Explain how solution designs and data-flow diagrams would make integration and operational responsibilities clear.

How would you establish architectural standards and guardrails across an AI delivery team?

What to coverPreparation suggestion: Describe how you would define reusable patterns, document decisions and review designs across models, pipelines and APIs. Address how guardrails can preserve security and maintainability without blocking delivery.

How would you decide whether to build, buy or integrate an AI platform component?

What to coverPreparation suggestion: Compare functional fit, integration effort, security, data governance, operational support and total cost. Explain how you would record the trade-offs and their effect on a technical roadmap.

How would you design an APAC deployment when data-sovereignty rules differ by country?

What to coverPreparation suggestion: Explain how you would identify where data may be stored and processed before choosing Azure, AWS or another compliant option. Discuss when multi-cloud or hybrid architecture is justified and what additional operational complexity it creates.

How would you balance cloud cost, reliability and reproducibility for AI workloads?

What to coverPreparation suggestion: Discuss workload sizing, cost visibility and the trade-off between shared resources and workload isolation. Cover repeatable environments, access control and the support needed to diagnose failures in development and deployment.

How would you prepare large, governed datasets for an AI use case?

What to coverPreparation suggestion: Explain how SQL and Python could support data discovery, quality checks and preparation on a lake, lakehouse or warehouse. Address lineage, access permissions and how quality findings would affect whether a dataset is ready for use.

How would you choose among a foundation model, fine-tuning and RAG for a use case?

What to coverPreparation suggestion: Compare the need for grounded information, adaptation, evaluation evidence, data controls, latency and cost. Note that the posting asks for use-case-specific selection rather than naming a preferred model or approach.

What would you include in an MLOps or LLMOps pipeline for an APAC AI service?

What to coverPreparation suggestion: Cover model and data versioning, evaluation before release, monitoring after deployment and criteria for retraining or reassessment. Explain how these controls support reproducibility and operational decisions.

How would you assess and mitigate security risks in a new AI architecture?

What to coverPreparation suggestion: Describe how you would examine data flows and access paths, then incorporate encryption, access controls and audit logging. Explain how you would work with security and risk stakeholders to document residual risks and align with global governance.

How would you set a regional AI platform strategy around shared services?

What to coverPreparation suggestion: Identify which capabilities might be reusable across deployments and which decisions may need to remain local for regulatory or operational reasons. Discuss integration patterns, ownership and how new model developments would prompt a roadmap review.

Behavioural practice

Tell me about a time you guided a delivery team through a difficult architecture decision.

What to coverPreparation suggestion: Use a real situation, your actions to clarify standards and trade-offs, and the resulting delivery or maintainability outcome. Distinguish your technical direction from decisions made collectively by the team.

Describe a time you translated a business challenge into a data or AI solution.

What to coverPreparation suggestion: Explain the situation, how you worked with business and technical stakeholders to define the problem, and what resulted. Include how you tested whether available data could support the proposed approach.

Give an example of collaborating across different APAC teams or stakeholder perspectives.

What to coverPreparation suggestion: Describe the situation, how you resolved differing needs or constraints, and the outcome. Focus on concrete communication and decision-making actions rather than general claims about collaboration.

Tell me about a time you raised a security, privacy or governance concern in an initiative.

What to coverPreparation suggestion: Set out the risk, the stakeholders you engaged and the mitigation you proposed. Explain the result, including any trade-off between delivery goals and risk reduction.

Describe a time you sponsored or shaped an AI, IT or automation initiative at scale.

What to coverPreparation suggestion: State the business context, your actions in weighing value, cost and risk, and the measurable or observable result. Be clear about your own role and any limits of the outcome.

Sharpen your knowledge

Topics to prepare

Architecture standards and platform strategy
The posting calls for technical guardrails across models, data pipelines, APIs, cloud infrastructure and security, plus shared services and reusable components. Revise by explaining which decisions should be standardized regionally and which should remain use-case-specific for the Marine Ground Digital Twin Platform.
APAC cloud and data sovereignty
Fugro specifies Azure, AWS and compliant options, with multi-cloud or hybrid designs where regulations demand them. Revise by comparing two hypothetical deployment designs against data-location constraints, reliability, operational support and cost; do not assume one design fits every APAC country.
Governed data platforms
The requirements cover lakes, lakehouses or warehouses; large-dataset preparation; and data quality and governance. Revise by tracing how a dataset moves from ingestion to governed use in analytics or AI, identifying quality checks, ownership and access decisions.
AI approach and lifecycle choices
The role evaluates foundation models, fine-tuning and RAG, and designs MLOps and LLMOps pipelines. Revise by comparing these approaches for a hypothetical use case, then specifying versioning, evaluation, monitoring and retraining triggers. The use case is a preparation exercise, not a stated Fugro project.
Security, privacy and technical risk
The posting explicitly requires security-by-design, privacy-by-design and access control, with encryption, audit logging and technical risk mitigation in the architecture. Revise by mapping a hypothetical AI data flow to sensitive-data risks, controls and residual risks.
Architecture leadership and business trade-offs
Fugro seeks technical leadership in collaborative delivery, stakeholder collaboration, and financial and risk governance. Revise by rehearsing how to explain an architecture trade-off to delivery, business, security and risk stakeholders, including its cost and maintainability implications.

Skills to revise

SQL and Python data preparation
Both are explicit requirements. On a practice dataset, use SQL and Python to inspect volume and missing values, identify duplicates, apply documented quality rules and prepare a governed analysis-ready output.
Architecture artefact production
The role produces solution designs, data flows, integration maps and technical roadmaps. Draft a one-page architecture for a hypothetical AI-enabled platform, showing data movement, API boundaries, access controls and operational ownership.
Build–buy–integrate assessment
Fugro expects the architect to evaluate platforms and vendors. Create a decision matrix for a hypothetical platform component, scoring integration effort, governance fit, security, reliability and ongoing cost; state the assumptions behind the recommendation.
AI pipeline design
The responsibilities specify experimentation, deployment and monitoring environments. Sketch a model lifecycle with versioned data and models, evaluation gates, monitoring signals and retraining triggers, then explain how results would be reproduced.
Control and risk review
The role embeds encryption, access controls and audit logging and conducts technical risk assessments. Annotate your practice architecture with each control, identify one remaining technical risk and propose a mitigation for review with security and risk stakeholders.

Your preparation checklist

Prepare an architecture walkthrough
Choose a relevant project, if available, and bring a one-page solution diagram showing data flows, APIs, cloud components, security boundaries, trade-offs and your architectural decisions. Relate it to Fugro’s stated Marine Ground Digital Twin Platform responsibility without assuming its implementation.
Practice a regional design exercise
Sketch a proposed AI architecture for APAC deployment, with an integration map and brief roadmap. Explain how you would choose between Azure, AWS and compliant alternatives, and when data-sovereignty rules might require hybrid or multi-cloud design.
Assemble governed-data evidence
Prepare a concrete example of using SQL and Python to discover, analyse and prepare large datasets. Show the data-quality checks, governance and access controls used, and how the work served an analytics or AI need.
Draft an AI approach comparison
Create a short decision matrix for foundation-model integration, fine-tuning and RAG, including build-versus-buy-versus-integrate criteria. These are approaches the role evaluates, not assumed components of the platform.
Map operational and security guardrails
Prepare a sample MLOps or LLMOps flow covering versioning, evaluation, monitoring and retraining triggers. Add encryption, audit logging, access control, reproducibility, reliability, operational support and cost trade-offs.
Rehearse technical-leadership examples
Select examples, if available, of setting engineering standards, guiding an agile delivery team, collaborating with business and risk stakeholders, and weighing financial and technical risks. State your decisions and their outcomes.
Bring focused employer questions
Ask about the Digital Twin Platform’s current architecture, APAC data-residency constraints, global governance standards, ownership of shared AI services and priorities for the three-year Singapore placement.

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…