New opportunity

Research Engineer I / Research Engineer II (Computer Science/Artificial Intelligence)

Nanyang Technological University · Singapore

About this role

The Alibaba-NTU Global e-Sustainability CorpLab (ANGEL) seeks a Research Engineer to develop PrimeNet. The lab develops green digital technologies and sustainable digital solutions through a collaboration between Alibaba Group and Nanyang Technological University.

Responsibilities

Research new algorithms, review relevant literature, and stay current with advances in the field.
Design experiments to test hypotheses and validate models; turn research concepts into working prototypes.
Develop software, algorithms, or hardware systems for real-world applications, write clean and scalable Python code, and collect, preprocess, and analyze large datasets.

Requirements

A bachelor’s or master’s degree in Computer Science, Artificial Intelligence, or a related field.
Demonstrated experience in natural language processing, machine learning, and deep learning; familiarity with TensorFlow or PyTorch.
Proficiency in Python for implementing NLP and generative AI algorithms, and an understanding of text processing and generation.
Ability to analyze complex problems and data, develop creative solutions to research challenges, and communicate and document findings for diverse audiences.

The position is full-time at NTU Main Campus, Singapore. No specific number of years of experience or salary is stated.

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

Skills for this role

PythonNatural language processingMachine learningDeep learningGenerative AITensorFlowPyTorchAlgorithm developmentExperimental designData preprocessingData analysisResearch communication

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

Algorithm research grounded in literature
The role calls for developing new algorithms while reviewing relevant literature. Prepare to explain how you identify a research gap and choose an approach worth testing.
Rigorous model validation
The posting emphasizes experiments that test hypotheses and validate models. Prepare an example that distinguishes the hypothesis, comparison method, results, and limitations.
NLP and generative AI implementation
Requirements explicitly cover NLP, machine learning, deep learning, and Python implementations for text processing and generation, with familiarity with TensorFlow or PyTorch. Focus preparation on explaining implementation choices rather than naming tools alone.
Working prototypes and data pipelines
Responsibilities span large-dataset collection, preprocessing, and analysis; clean, efficient, scalable Python; and turning research into prototypes for real-world applications. Prepare to describe how data handling and code design supported a usable prototype.
Communicating research findings
The role requires documenting findings and presenting them to diverse audiences. Prepare a concise explanation of a technical result, including what the evidence does and does not establish.

Likely challenges to think through

Balancing novelty with credible evidence
A likely challenge, not a stated employer problem, is finding a useful new algorithm without mistaking an apparent improvement for a validated result. Prepare to discuss baselines, controlled comparisons, and how you would interpret inconclusive findings.
Balancing rapid prototypes with scalable code
Because the role combines research prototypes, large datasets, and efficient Python, a likely trade-off is speed of iteration versus maintainability and performance. Prepare an example of how you would prioritize a testable prototype, then improve its data pipeline or code as evidence warrants.
Scoping work when PrimeNet details are unspecified
The posting names PrimeNet and ANGEL’s sustainability mission but does not describe PrimeNet’s architecture, datasets, or success measures. A likely preparation challenge is avoiding unsupported assumptions: be ready to explain which technical and application constraints you would clarify before selecting an NLP approach or evaluation metric.

Practice your answers

Technical practice

How would you use recent research literature to develop a new algorithm?

What to coverThe role involves reviewing literature and developing algorithms. Suggested preparation: explain how you would compare relevant methods, identify a testable gap, and decide whether a new approach improves on a baseline.

How would you prepare a large text dataset for an NLP experiment?

What to coverThe posting calls for collecting, preprocessing, and analyzing large datasets. Suggested preparation: discuss text cleaning, data quality, and train–test separation, including how preprocessing choices could affect results.

What choices matter when representing text for a machine-learning model?

What to coverThe role requires understanding NLP concepts and text processing. Suggested preparation: compare plausible text representations or tokenization choices and explain how you would assess their effect on the intended task.

How would you choose an approach for a text-generation task?

What to coverThe posting specifies machine learning, deep learning, and Python implementations of generative AI algorithms. Suggested preparation: compare a suitable baseline with a more complex approach, addressing output quality, data needs, and computational cost without assuming PrimeNet has a particular architecture.

How would you design an experiment to validate a model?

What to coverDesigning experiments to test hypotheses and validate models is a stated responsibility. Suggested preparation: define a hypothesis, baseline, evaluation measures, and controls that would make the comparison interpretable.

How would you investigate an NLP model that performs well overall but fails on some inputs?

What to coverThe role calls for analyzing complex problems and deriving insights from data. Suggested preparation: describe how you would inspect failure patterns, check data and preprocessing, and use targeted tests before changing the model.

How would you implement and debug a deep-learning experiment in TensorFlow or PyTorch?

What to coverThe posting asks for familiarity with frameworks such as TensorFlow or PyTorch. Suggested preparation: walk through data loading, training, validation, and checks for implementation errors; use whichever framework you know rather than implying both are required.

How would you make Python code for a research prototype clean and scalable?

What to coverWriting clean, efficient, and scalable Python code is a stated responsibility. Suggested preparation: discuss modularity, tests, profiling, and trade-offs between quick experimentation and maintainable implementation.

How would you turn a research concept into a working prototype?

What to coverThe role includes translating research into prototypes and developing systems for real-world applications. Suggested preparation: outline a minimal implementation, how you would verify its behavior, and what evidence would justify further development.

How would you assess computational cost alongside model performance?

What to coverANGEL's stated mission concerns green digital technologies, while the role involves validating models. Suggested preparation: consider reporting resource use alongside task results and explain the quality–cost trade-off; the posting does not specify a required sustainability metric for PrimeNet.

Behavioural practice

Tell me about a research challenge that required a creative solution.

What to coverThe posting calls for creative approaches to research challenges. Suggested preparation: use a real example that sets out the situation, the alternative you tried and why, and the result or lesson.

Describe a time an experiment did not support your hypothesis.

What to coverTesting hypotheses and validating models are stated responsibilities. Suggested preparation: explain the situation, how you checked the experiment or revised your approach, and what the result showed.

Tell me about turning a research idea into a working prototype.

What to coverTranslating research concepts into prototypes is part of the role. Suggested preparation: describe the initial situation, your implementation and testing decisions, and the resulting evidence about whether the idea was workable.

Describe how you communicated technical findings to different audiences.

What to coverThe posting asks for written and verbal communication of research findings to diverse audiences. Suggested preparation: give a real situation, explain how you adapted the message or documentation, and state the outcome.

Tell me about using new literature to change your technical approach.

What to coverThe role involves exploring literature and staying current with advances. Suggested preparation: describe what prompted the review, how you assessed and applied a finding, and what resulted.

Sharpen your knowledge

Topics to prepare

NLP for text processing and generation
Review how text is represented, processed, and generated; explain how those choices affect model behavior. The posting explicitly requires NLP experience and understanding of text processing and generation.
Machine learning and deep learning
Compare two model approaches for a text task, including what each learns and how you would assess its results. The posting explicitly requires experience in machine learning and deep learning.
Generative AI algorithms
Walk through a generative text model’s inputs, training objective, outputs, and likely failure cases. This supports the stated requirement to implement NLP and generative AI algorithms in Python.
Research hypotheses and model validation
Take a proposed algorithm change and specify a testable hypothesis, a comparison, and evidence that would support or reject it. The role explicitly involves developing algorithms and designing experiments to validate models.
Research literature and application context
Summarize one relevant NLP research paper: its contribution, evidence, limitations, and a prototype idea. Literature review and prototyping are stated responsibilities. Relating the idea to sustainability is a preparation suggestion based on ANGEL’s mission, not a stated PrimeNet specification.

Skills to revise

Python NLP implementation
Write a small text-processing and generation prototype in clean Python, then explain how you would improve its efficiency and scalability. The posting explicitly calls for Python proficiency and clean, efficient, scalable code.
TensorFlow or PyTorch practice
Use either framework to train and evaluate a small deep-learning text model, and explain the training and inference steps. The requirement is familiarity with frameworks such as TensorFlow or PyTorch, not both.
Dataset preparation and analysis
Prepare a text dataset, document preprocessing decisions, and inspect patterns or errors in the resulting data. Collecting, preprocessing, and analyzing large datasets are stated responsibilities.
Experimental reporting
Run a controlled comparison for a model change and produce a brief report covering the hypothesis, results, limitations, and next step; practice explaining it to a non-specialist. This reflects the stated experiment-design and research-communication responsibilities.

Your preparation checklist

Assemble a requirements evidence sheet
List verifiable examples of the stated degree, NLP, machine learning, deep learning, Python, and TensorFlow or PyTorch requirements. Note where each example can be shown or explained.
Prepare an NLP project walkthrough
Select a real project and outline the problem, your contribution, text-processing or generation approach, model choices, results, and limitations. Include a relevant Python code sample if available.
Diagram a large-data workflow
Show how you collected, preprocessed, checked, and analyzed a dataset in a real project. Be ready to explain choices affecting data quality and efficient, scalable Python code.
Write a short research comparison
Compare relevant NLP or generative AI approaches from literature you know, identify an open question, and propose an algorithmic idea. Present it as preparation, not as a claim about PrimeNet’s design.
Draft an experiment plan
For a proposed text-processing or generation task, state a hypothesis, baseline, evaluation measures, and how you would examine errors and validate the model. The posting does not specify PrimeNet’s tasks or metrics.
Sketch a prototype design
Prepare a simple design showing how a research idea could become a working software prototype, including inputs, outputs, components, and tests. Explain the trade-offs you would revisit for a real-world application.
Practice a research explanation
Prepare a brief written summary and spoken explanation of one finding for both technical and nontechnical audiences, including what the evidence does and does not support.
Bring questions about the role
Ask what PrimeNet is intended to do, which datasets and evaluation criteria matter, how its work relates to ANGEL’s sustainability mission, and how responsibilities differ between Research Engineer I and II.

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…