New opportunity

Senior Manager, Artificial Intelligence

IMD Info-communications Media Development Authority · Singapore

About this role

The Senior Manager, Artificial Intelligence, works within IMDA’s Business Technology Group to develop AI solutions, particularly generative AI, and support responsible AI adoption across Singapore. The role contributes to AI initiatives and ecosystem development in sectors including legal and healthcare.

Responsibilities

Assess AI technology providers and interpret technical articles and research to provide AI expertise within the organisation.
Lead generative AI solution development across software and hardware, working with technical and domain experts to identify use cases and build early prototypes.
Contribute to AI hardware benchmarking, agentic AI and embodied AI projects. Integrate safety and trust considerations into AI development.
Design and eventually run grant programmes for end-user AI adoption. Support programmes covering providers’ research and commercialisation and AI workforce development.
Facilitate AI workshops, events and overseas trips that support knowledge exchange and collaboration.

Requirements

A master’s degree in Computer Science, Artificial Intelligence, Machine Learning or a related field; a PhD is preferred.
At least a few years of hands-on AI development experience, including building and deploying solutions in real-world settings. Strong knowledge of generative AI principles and development practices, AI safety and trust, and applications across sectors is required.
Experience partnering with local research ecosystems and government agencies on technology projects; strong analytical, technical due diligence and communication skills; proficiency in an AI programming language such as Python; and familiarity with AI frameworks and tools.

Preferred qualifications include experience developing technology-sector grant programmes and hands-on work in AI hardware benchmarking, agentic AI or embodied AI. The position is full-time in Singapore. Only shortlisted applicants will be notified; the position will commensurate with experience.

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

Skills for this role

Generative AIAI developmentAI safetyTechnical due diligenceAI hardware benchmarkingAgentic AIEmbodied AIPrototypingGrant program developmentPythonResearch analysisCommunicationStakeholder collaboration

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

Lead generative AI development
The role leads software- and hardware-related solution development with technical and domain experts, from identifying use cases to building early prototypes. The posting also calls for experience deploying AI in real-world settings.
Provide technical due diligence
Assess AI technology providers, interpret technical articles and research, and turn the findings into usable AI expertise for the organisation.
Build in safety and trust
Incorporate AI safety and trust considerations into development processes, alongside the role’s focus on responsible AI adoption across sectors such as legal and healthcare.
Shape AI adoption grants
Craft and eventually run grant programmes for end-user AI adoption, while supporting programmes for providers’ research and commercialisation and AI workforce development.
Connect emerging AI work and partners
Contribute to AI hardware benchmarking, agentic AI and embodied AI initiatives. Work with researchers, industry, government agencies and other stakeholders, including through workshops and knowledge-exchange events.

Likely challenges to think through

Moving from prototype to useful adoption
Preparation hypothesis: early generative AI use cases may need to demonstrate practical value without being treated as deployment-ready. Prepare an example of how you would choose a use case, test a prototype with domain experts and decide what evidence is needed before wider use.
Balancing technical promise with trust
Preparation hypothesis: a promising model or provider may perform well technically yet raise safety or reliability concerns. Prepare to explain how you would compare evidence, identify limitations and incorporate safeguards before recommending its use.
Serving different parts of the AI ecosystem
Preparation hypothesis: end-user adoption grants and programmes supporting research, commercialisation and workforce development may call for different measures of success. Prepare a reasoned approach to defining programme objectives and communicating trade-offs to technical and non-technical partners.

Practice your answers

Technical practice

How would you choose a generative AI approach for a new use case?

What to coverSuggested answer focus: explain how the task, available data, quality needs and practical constraints shape the approach. Compare options through a small evaluation rather than assuming one model or development method fits every use case.

How would you turn a domain expert’s idea into an early AI prototype?

What to coverSuggested answer focus: define the user problem and success criteria with technical and domain experts, then describe a limited prototype and how you would test it. Address what evidence would justify further development.

What would you check when a deployed AI solution underperforms?

What to coverSuggested answer focus: distinguish problems with inputs, model behaviour, integration and operating conditions. Explain how you would isolate the cause, measure any improvement and decide whether to revise or withdraw the solution.

How would you conduct technical due diligence on an AI provider?

What to coverSuggested answer focus: identify which technical claims need independent evidence and how you would test relevance to the intended use case. Explain how you would communicate limitations and unresolved questions to decision-makers.

How would you interpret a research article for an organisational decision?

What to coverSuggested answer focus: separate the article’s demonstrated results from its assumptions and broader claims. Describe what additional validation would be needed before applying its findings.

How would you design a useful AI hardware benchmark?

What to coverSuggested answer focus: choose representative workloads and comparable measurements for the intended AI application. Discuss trade-offs among performance, resource use and reproducibility, while avoiding conclusions beyond what the benchmark tests.

How would you evaluate an agentic AI prototype?

What to coverSuggested answer focus: define the tasks and boundaries within which the system may act, then assess task completion and failure modes. Include how human oversight and safety checks would affect a decision to proceed.

What would you examine in an embodied AI proof of concept?

What to coverSuggested answer focus: explain how you would test the interaction between the AI system, its inputs and its physical environment. Discuss reliability and safety limits before considering a wider trial.

How would you build safety and trust into a generative AI solution?

What to coverSuggested answer focus: identify risks associated with the use case and incorporate evaluation, safeguards and human review during development. Explain how you would monitor limitations after deployment rather than treating safety as a final check.

How would you design an AI-adoption grant programme for end users?

What to coverSuggested answer focus: distinguish the programme’s adoption goal from programmes supporting provider research, commercialisation or workforce development. Propose clear objectives and ways to assess outcomes without assuming any existing grant rules.

Behavioural practice

Tell me about a time you led a prototype with technical and domain experts.

What to coverSuggested practice: use a real situation, your actions and the result to show how you aligned people on the use case and scope. Explain how differing expert views affected the prototype.

Describe a technology project involving research or public-sector partners.

What to coverSuggested practice: give the situation, your actions in coordinating the partners and the result. Focus on how you handled differing objectives or evidence needs without overstating your authority.

Tell me about a time you explained complex AI findings to different audiences.

What to coverSuggested practice: describe the situation, how you adapted your explanation for technical and non-technical listeners, and the result. Show how you preserved important uncertainty or limitations.

Describe a time a safety concern changed an AI development decision.

What to coverSuggested practice: set out the situation, the concern, your actions and the result. Explain the trade-off you made between usefulness and responsible implementation.

Tell me about a workshop or knowledge-sharing event you helped deliver.

What to coverSuggested practice: describe its purpose, your actions in bringing people together and the result. Connect what participants learned or contributed to a concrete next step, if there was one.

Sharpen your knowledge

Topics to prepare

Generative AI principles and development choices
Compare two approaches to a generative AI use case, explaining their technical trade-offs and what evidence would justify a prototype. The posting requires strong generative AI literacy and leadership of solution development across software and hardware.
AI safety and trust
For a proposed AI use case, identify plausible failures, affected users and checks needed before deployment. The role explicitly calls for safety and trust considerations throughout development.
Sector-specific AI adoption
Contrast a legal and a healthcare use case: identify the domain expertise needed, a useful outcome and a reason each might not be suitable for an early prototype. The posting names both sectors and calls for collaboration with domain experts.
Research and provider assessment
Read a technical AI article and separate its claims, supporting evidence and unanswered questions for a potential provider assessment. Interpreting research and conducting technical due diligence are central duties.
AI ecosystem and adoption programmes
Sketch how an end-user adoption programme would differ in purpose from support for provider research, commercialisation or workforce development. The role will help craft and eventually run grants and support these other programmes.
Hardware benchmarking, agentic AI and embodied AI
Outline what each area would need to demonstrate in an early project, without assuming a particular product or tool. These are named initiatives; hands-on experience in any of them is described as highly desirable, not required.

Skills to revise

Build and explain an AI prototype
As preparation, implement a small generative AI use case in Python or another AI programming language, then document how it could be deployed, tested and reviewed for safety. The posting asks for hands-on development and real-world deployment experience, Python proficiency and familiarity with AI frameworks and tools.
Produce a technical due-diligence brief
Create a one-page comparison of two hypothetical AI providers, stating evidence to request, technical uncertainties and a recommendation with caveats. This rehearses the advertised provider-assessment duty without assuming IMDA uses a particular assessment process.
Design a hardware benchmark
As optional preparation, draft a repeatable comparison plan for an AI workload across two hardware options, including measurements and limitations. AI hardware benchmarking is an ongoing initiative, and hands-on experience is highly desirable.
Facilitate a cross-disciplinary discussion
Practise a short use-case workshop agenda and explain its technical choices once to an AI specialist and once to a non-technical domain expert. The role works with technical and domain experts, partners with research and government stakeholders, and supports workshops.
Draft an adoption-programme outline
As preparation, write a brief outline for an end-user AI adoption grant covering its intended outcome, what applicants would demonstrate and how results could be reviewed. The role will craft and eventually run such programmes; grant-development experience is a plus.

Your preparation checklist

Map the stated requirements to evidence
Prepare a one-page inventory of your degree, AI development and deployment work, Python and framework experience, safety work, and any relevant sector experience. Mark gaps for discussion rather than implying experience you do not have.
Rehearse a deployed AI project walkthrough
Choose a real project and explain the use case, your leadership and hands-on contributions, technical choices, deployment, measured results, and what you would change. Emphasise generative AI where your experience supports it.
Draft a technical due-diligence brief
Use a provider or research article you already know to practise summarising its claims, evidence, limitations and questions for verification. This mirrors the role’s provider assessment and research-interpretation work.
Sketch an early generative AI prototype
As a practice exercise, outline a legal or healthcare use case: what to validate with domain experts, a software and hardware approach, prototype tests, and go/no-go criteria. Label the design as hypothetical.
Prepare a safety and benchmarking discussion
Show how you would test reliability, misuse and other safety or trust risks during development. Bring actual hardware-benchmarking, agentic AI or embodied AI evidence if available; otherwise, discuss a proposed evaluation approach without claiming hands-on work.
Document a partnership example
Prepare a concise account of any technology project with researchers or a government agency, covering stakeholder roles, decisions and outcomes. The posting specifically seeks experience partnering with local research ecosystems and government agencies.
Outline an AI adoption grant concept
Draft a proposed end-user adoption programme with its purpose, possible assessment measures and how outcomes might be monitored. Treat these as interview ideas, not as IMDA’s existing grant rules.
Practise facilitation and prepare employer questions
Explain one complex AI concept in technical and non-technical terms, as you might in a workshop. Bring questions about the team’s near-term AI initiatives, how domain experts shape prototypes, and the scope of its grant-programme work.

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…