New opportunity

Staff Data Scientist, Algorithm (Risk Product – AML & Financial Crime)

Airwallex · Singapore

About this role

As a Staff Data Scientist in the Risk Product Data Science team, you will be a technical leader responsible for advancing how Airwallex proactively identifies and mitigates financial crime risk. You will work across customer, transaction, behavioral, network and external intelligence data to detect known and emerging risks, uncover suspicious communities and coordinated activity, and improve how we screen customers and counterparties for financial crime exposure. A key part of the role is developing AI-powered risk detection capabilities, including name screening, entity resolution, network intelligence, and proactive discovery of previously unknown risk patterns. You will partner closely with Risk Product, Engineering, Financial Crime Compliance and Risk Operations to translate complex risk problems into scalable, production-grade controls. Responsibilities: Lead the Data Science strategy for AML and Financial Crime risk controls across Risk Product. Develop proactive risk detection to uncover emerging threats, suspicious behaviors, control gaps and previously unknown financial crime patterns. Develop and scale advanced risk capabilities across AI-powered screening, graph and network intelligence, entity resolution, and machine learning. Advance key AML/FCC use cases including name screening, transaction monitoring, and detection of coordinated activity and fraud or financial crime rings. Partner with Risk Product, Engineering, Financial Crime Compliance and Risk Operations to translate complex risk problems into effective, scalable production controls. Influence the Risk Product roadmap and raise the technical bar for Risk Data Science through technical leadership, reusable frameworks, and mentoring. Minimum Requirements: 7+ years of experience in Data Science, Machine Learning, Applied AI, Risk Analytics, or a related quantitative field, with demonstrated impact at Staff, Lead, or equivalent scope. Strong expertise in applied machine learning and modern AI, with experience building and productionizing data-driven solutions at scale. Experience with one or more areas highly relevant to financial crime detection, such as NLP/LLMs, entity resolution, information retrieval, graph or network analytics, anomaly detection, or risk scoring. Strong SQL and Python skills are must. Strong communication and stakeholder management skills. Preferred Qualifications: Experience in fintech, payments, banking, marketplaces or another regulated or high-risk domain. Experience in AML, financial crime, sanctions, PEP screening, adverse media, fraud, trust and safety or transaction monitoring. Familiarity with modern LLM evaluation, prompt/model experimentation, retrieval quality assessment and AI observability. Experience applying graph analytics to detect mule networks, collusive behavior, coordinated account activity or interconnected high-risk entities.

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

Skills for this role

Machine LearningApplied AIRisk AnalyticsNLPLLMsEntity ResolutionInformation RetrievalGraph AnalyticsNetwork AnalyticsAnomaly DetectionRisk ScoringSQLPythonAMLFinancial Crime ComplianceTransaction MonitoringAI Observability

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

Proactive Risk Discovery Strategy
The role requires moving beyond traditional rule-based systems to proactively identify emerging financial crime patterns. You must demonstrate how to design and implement discovery frameworks that uncover unknown threats, rather than just optimizing existing supervised models.
Scalable Production-Grade AI Integration
You are expected to translate complex risk problems into scalable, production-grade controls. This requires balancing advanced AI techniques—such as graph analytics, entity resolution, or LLMs—with the operational constraints of a high-volume, regulated fintech environment.
Cross-Functional Stakeholder Influence
As a technical leader, you must effectively partner with Product, Engineering, Compliance, and Risk Operations. Success depends on your ability to communicate complex technical risk strategies to non-technical stakeholders and align them with the broader Risk Product roadmap.

Likely challenges to think through

Balancing Innovation with Regulatory Rigor
Implementing cutting-edge AI (like LLMs or graph-based anomaly detection) in a highly regulated AML environment creates a tension between model explainability and performance. Prepare to discuss how you maintain auditability and compliance while deploying advanced, potentially 'black-box' algorithms.
Data Complexity and Entity Resolution
Financial crime detection relies on disparate data sources, including behavioral, network, and external intelligence. The challenge lies in performing accurate entity resolution at scale to prevent false positives while maintaining high recall for sophisticated, coordinated criminal activity.

Practice your answers

Technical practice

Graph Analytics for Financial Crime Detection

What to coverExplain how you would utilize graph databases or network analysis to identify mule networks or collusive behavior. Discuss the trade-offs between using community detection algorithms versus path-finding heuristics in a high-volume payment environment.

Entity Resolution at Scale

What to coverDescribe your approach to building a robust entity resolution system for name screening. Address how you handle variations in data quality, such as misspellings or partial matches, and how you balance precision and recall to minimize false positives for compliance teams.

Proactive Risk Discovery vs. Supervised Learning

What to coverThe role emphasizes proactive discovery of unknown risk patterns. Contrast this with traditional supervised learning for fraud detection and explain how you would design an unsupervised or semi-supervised framework to surface emerging threats.

LLM Evaluation in Risk Systems

What to coverGiven the focus on modern AI, how would you implement an evaluation framework for LLMs used in adverse media screening? Focus on metrics for retrieval quality, hallucination mitigation, and the importance of AI observability in a regulated environment.

Productionizing Machine Learning Models

What to coverDiscuss the challenges of deploying machine learning models into a real-time payment processing pipeline. Address how you ensure low-latency inference, model monitoring, and the feedback loop required to maintain performance as financial crime tactics evolve.

Feature Engineering for Transaction Monitoring

What to coverWhat specific behavioral features would you prioritize to detect coordinated account activity? Explain how you would aggregate and transform raw transaction data to capture temporal patterns indicative of money laundering.

SQL and Data Architecture for Risk

What to coverDescribe how you structure complex SQL queries to perform large-scale risk analysis across disparate datasets. Suggest how you would optimize data pipelines to ensure that Risk Operations has timely access to actionable intelligence.

Handling Imbalanced Datasets in Fraud

What to coverFinancial crime data is notoriously imbalanced. Discuss your preferred techniques for training models when the positive class (fraudulent activity) is extremely rare compared to legitimate transactions.

Behavioural practice

Influencing Cross-Functional Stakeholders

What to coverDescribe a situation where you had to influence Product, Engineering, or Compliance teams to adopt a new data science strategy. Focus on how you translated complex technical risk problems into business value to gain alignment.

Mentorship and Technical Leadership

What to coverProvide an example of how you have raised the technical bar within a data science team. Discuss how you mentored junior members or introduced reusable frameworks that improved the team's overall output.

Navigating Regulatory Constraints

What to coverTell us about a time you had to balance the need for innovative AI solutions with strict regulatory or compliance requirements. How did you ensure your model was both effective and explainable to auditors?

Managing Ambiguity in Risk Discovery

What to coverDescribe a project where you were tasked with uncovering a previously unknown risk pattern. How did you define the scope, identify the necessary data sources, and validate your findings when no ground truth existed?

Sharpen your knowledge

Topics to prepare

Financial Crime Risk Frameworks
Review AML, sanctions, and PEP screening regulatory requirements. Exercise: Map out a high-level architecture for a transaction monitoring system that balances regulatory compliance with low-latency production performance, specifically addressing how to handle false positives.
Graph Analytics for Fraud Detection
Study methodologies for identifying mule networks and collusive behavior. Exercise: Design a graph-based approach to detect circular transaction patterns or coordinated account activity, focusing on how to represent entities and relationships in a scalable database.
LLM Evaluation and AI Observability
Focus on modern LLM deployment, including prompt experimentation and retrieval quality assessment. Exercise: Define a set of metrics to evaluate the performance of an LLM-based adverse media screening tool, specifically focusing on precision, recall, and drift detection in a production environment.

Skills to revise

Advanced SQL and Python for Risk
Practice writing complex SQL queries for feature engineering on large-scale transaction datasets. Exercise: Write a Python script to perform entity resolution on noisy customer data, utilizing fuzzy matching or blocking techniques to identify duplicate records.
Proactive Risk Discovery
Develop techniques for unsupervised anomaly detection to find unknown risk patterns. Exercise: Implement an isolation forest or similar anomaly detection algorithm on a synthetic transaction dataset to identify outliers that deviate from established behavioral baselines.

Your preparation checklist

Map your experience to AML and financial crime use cases
Prepare to discuss specific instances where you applied machine learning to detect money laundering, fraud, or sanctions violations. Focus on your ability to move beyond simple rule-based systems to proactive discovery of emerging risk patterns.
Develop a technical framework for graph and network analytics
Create a high-level architectural overview of how you would use graph analytics to identify mule networks or collusive behavior. Be ready to explain your choice of tools and how you handle large-scale, interconnected entity data.
Prepare a case study on productionizing AI models
Select a past project where you took a model from research to production. Emphasize your experience with AI observability, model evaluation, and how you ensured the solution was scalable and reliable for high-stakes risk environments.
Refine your approach to entity resolution and name screening
Review your experience with information retrieval and NLP techniques used for entity resolution. Be prepared to discuss how you balance precision and recall when screening customers against sanctions or PEP lists.
Formulate questions for cross-functional stakeholders
Prepare thoughtful questions about how the Data Science team currently collaborates with Financial Crime Compliance and Risk Operations. Focus on how technical requirements are translated into production controls and how the team measures the effectiveness of their risk models.

Availability last checked: 18 Oct 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…