New opportunity

AI Engineer

Ethos BeathChapman · Singapore

About this role

Ethos BeathChapman is recruiting an AI Engineer for a professional services client building its AI and data capability. The role focuses on delivering AI solutions that address business problems, improve internal processes and reduce manual work.

Responsibilities

Build and deploy AI solutions; design, fine-tune and evaluate LLMs and other machine-learning models using internal data in a secure infrastructure environment.
Integrate AI capabilities into existing internal systems and assess where commercial AI tools are appropriate.
Explore practical uses for open-source and commercial agentic AI tools.
Work with process owners, IT and governance stakeholders to ensure initiatives are secure, scalable and aligned with business needs.
Mentor junior team members and help develop the team's broader AI capability.

Requirements

A degree in Data Science, Computer Science, Business Analytics, Statistics, Mathematics or a related field.
At least 3 years of experience building and deploying AI/ML models in production, including AI agents.
Python programming or scripting ability; familiarity with ML frameworks and libraries such as PyTorch, TensorFlow, LangChain or LangGraph; and some exposure to cloud platforms such as Azure.
Understanding of data security, governance and model evaluation practices.
Strong written and verbal communication, attention to detail, self-motivation, an ability to meet deadlines and a willingness to learn.

The application requests Singapore employment status, including whether the applicant requires a valid Singapore work pass.

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

Skills for this role

PythonPyTorchTensorFlowLangChainLangGraphAzureLarge Language ModelsAI agentsModel fine-tuningModel evaluationData securityData governanceCloud platformsStakeholder collaborationMentoring

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

Deliver AI that improves real workflows
The role is to build and deploy AI solutions for business problems and internal process improvement, including reducing manual work. Prepare to explain how a production solution changed a workflow and how its value was measured.
Develop and evaluate models securely
The posting calls for designing, fine-tuning and evaluating LLMs and other ML models on internal data within secure infrastructure. Prepare to discuss model evaluation alongside data-security and governance decisions.
Integrate and select practical AI tools
The engineer will embed AI in existing systems, assess where commercial tools make sense, and explore practical uses for open-source and commercial agentic tools. Prepare to explain integration and tool-selection trade-offs rather than focusing on a single framework.
Coordinate delivery and build team capability
The role partners with process owners, IT and governance stakeholders to keep initiatives aligned, secure and scalable. It also includes mentoring junior colleagues and helping shape the team's AI capability.

Likely challenges to think through

Proving value beyond a working model
Preparation hypothesis: A technically sound model may not remove manual work if it does not fit the process. Prepare an example that connects a business problem to deployment, workflow adoption and a measurable outcome.
Balancing model utility with internal-data controls
Preparation hypothesis: Fine-tuning or evaluating on internal data may create tension between model performance and security or governance constraints. Prepare to explain how you would choose evaluation criteria and work with IT and governance stakeholders on acceptable data handling.
Choosing between custom and commercial approaches
Preparation hypothesis: Integration needs, scalability and control may pull in different directions when comparing an in-house model, an open-source agentic tool and a commercial product. Prepare a decision framework grounded in the existing system and the business use case.

Practice your answers

Technical practice

How would you decide whether an internal process needs an AI solution?

What to coverSuggested preparation: Explain how you would define the business problem, measure manual effort and establish a useful outcome with the process owner. Discuss when a simpler process change might be preferable to building a model.

How would you prepare internal data for fine-tuning an LLM or another ML model?

What to coverSuggested preparation: Cover data suitability, quality checks and access controls within a secure infrastructure setup. Explain how you would avoid using sensitive or poorly governed data inappropriately.

How would you evaluate an AI model before using it in an internal workflow?

What to coverSuggested preparation: Describe evaluation criteria tied to the workflow, representative test data and checks for failure cases. Explain how you would weigh model quality against operational cost and risk.

What would you check when a deployed AI solution stops meeting its intended business need?

What to coverSuggested preparation: Walk through how you would distinguish changes in input data, model behaviour and the surrounding workflow. Explain how monitoring and feedback could inform a fix or a decision to pause use.

How would you embed an AI capability into an existing internal system?

What to coverSuggested preparation: Discuss the system's inputs, outputs, access boundaries and how users would handle uncertain results. Address maintainability and the trade-off between a tightly integrated feature and a separate service.

How would you design a practical AI-agent workflow?

What to coverSuggested preparation: Explain what the agent should be allowed to do, how its actions would be constrained and when a person should review an outcome. Describe how you would test whether the agent improves the process rather than adding complexity.

How would you compare open-source and commercial AI tools for an internal use case?

What to coverSuggested preparation: Compare fit for the task, data handling, integration effort, cost and ongoing support. Explain how you would test a promising option before recommending its use.

How would you use Python and ML libraries to make model work reproducible?

What to coverSuggested preparation: Discuss clear data and evaluation code, versioning of inputs and model configurations, and repeatable tests. You could draw on a relevant framework named in the posting, such as PyTorch, TensorFlow, LangChain or LangGraph, without assuming one is mandated.

What would you consider when running AI workloads on secure cloud infrastructure?

What to coverSuggested preparation: Address access, data movement, scalability and operational cost, drawing on Azure or an equivalent platform if relevant. Keep the answer at the architecture and controls level; the posting does not specify particular cloud services.

Behavioural practice

Tell me about a time you aligned an AI initiative with process owners, IT and governance stakeholders.

What to coverSuggested preparation: Give the situation and each group's concern, then explain what you did to reach a workable decision. State the result and how you knew the solution still met the business need.

Tell me about a time you mentored a junior teammate on AI or ML work.

What to coverSuggested preparation: Describe the teammate's task and the support you provided without taking over their work. Explain the result for the teammate or the team's capability.

Tell me about a time you delivered technical work under a difficult deadline.

What to coverSuggested preparation: Set out the deadline and competing priorities, then describe how you chose what to complete and communicated any trade-offs. Give the result, including any limitations you made visible.

Tell me about a time you learned an unfamiliar AI tool or approach for a practical problem.

What to coverSuggested preparation: Explain the problem, how you assessed whether the new approach was useful and what you did to learn it. Describe the outcome, including if you decided not to adopt it.

Tell me about a time you explained a model limitation to a nontechnical stakeholder.

What to coverSuggested preparation: Describe the decision they needed to make and how you communicated the limitation in relevant terms. Explain their response and the resulting decision or change.

Sharpen your knowledge

Topics to prepare

Model evaluation
The posting calls for evaluating LLMs and other ML models. As preparation, take a hypothetical internal workflow and define a baseline, test cases, success measures and a review process for errors before deployment.
Data security and governance
The role uses internal data in a secure infrastructure setup and involves governance stakeholders. As preparation, map how hypothetical internal data would enter a model workflow, who could access it, and what approvals or controls you would discuss with governance.
Agentic and commercial-tool tradeoffs
The posting mentions production AI agents and assessing open-source and commercial agentic tools. As preparation, compare an agent-based approach with a simpler workflow for one manual task, including when a commercial tool would or would not make sense.
Production integration and scalability
The role involves embedding AI in existing internal systems and keeping initiatives scalable; it also asks for some cloud exposure, such as Azure or equivalent. As preparation, sketch a deployment design showing system boundaries, failure handling and how usage might grow.

Skills to revise

Python implementation
Solid Python programming or scripting ability is requested. As preparation, write a small script that processes fictional internal-workflow records, calls a mock model interface and handles invalid inputs; review it for clarity and testability.
Model fine-tuning and testing
The role includes designing, fine-tuning and evaluating LLMs and other ML models. As preparation, outline a small experiment using non-sensitive sample data: define the task, separate training and evaluation examples, and explain how you would compare results with a baseline.
Agent workflow prototyping
Production experience with AI agents is requested, and LangChain or LangGraph are examples of relevant libraries. As preparation, prototype or diagram an agent for a bounded internal task, showing its steps, stopping conditions and how a person would inspect its output.
Internal-system integration
Embedding AI capability into existing systems is an advertised responsibility. As preparation, draft an interface for a hypothetical internal application to request an AI result, including inputs, outputs and what happens when the model is unavailable.
Stakeholder explanation and mentoring
The role works with process owners, IT and governance stakeholders and mentors junior colleagues. As preparation, practise a short explanation of one proposed AI workflow for each audience, then outline how you would review a junior teammate’s implementation.

Your preparation checklist

Map your evidence to the requirements
Prepare a one-page, factual map of your degree, production AI/ML experience, Python work, relevant frameworks, cloud exposure and AI-agent experience against the posting. Note any gaps to discuss honestly.
Prepare a production project walkthrough
If you have a relevant example, outline the business problem, your contribution, model or agent design, deployment, results and lessons learned. The role calls for AI solutions that reduce manual work and improve internal processes.
Sketch a secure internal-data solution
As interview practice, draw a simple architecture for an LLM solution using internal data. Explain access controls, integration with an existing system, scalability and when fine-tuning would be appropriate.
Draft an evaluation plan
Prepare a sample plan covering a baseline, representative test cases, quality measures, failure analysis and governance review. Label it as a proposed approach, not a process used by this employer.
Compare agentic and commercial-tool options
Make a short decision table for an open-source agentic approach versus a commercial AI tool, covering practical use, integration effort, security and evaluation. The role explicitly explores both options.
Rehearse collaboration and mentoring examples
Prepare concise, truthful examples of working with process owners, IT or governance stakeholders, communicating trade-offs, meeting deadlines and helping junior colleagues, where your experience supports them.
Write targeted questions for the employer
Ask which internal processes are priorities, what secure infrastructure and governance constraints apply, how deployed models are evaluated, and how the AI Engineer will contribute to the developing team.

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…