SENIOR GRAPH DATA SCIENTIST / ML LEAD – FRAUD & FORENSICS
Level: Senior / Lead
CapCircle is recruiting a highly experienced Senior Graph Data Scientist / ML Lead for a fintech fraud detection and model modernisation programme.
This is a hands-on leadership role for a specialist with strong graph analytics, Python machine learning and fraud detection experience. The successful candidate will lead the graph and feature-engineering workstream, own the graph feature engine and drive the enhancement of an existing social engineering fraud model.
ROLE OBJECTIVE
Enhance the fraud detection model by integrating graph analytics, behavioural features and historical interaction data into a single feature set that improves precision and recall while reducing false positives. The solution must be explainable, auditable, scalable and reusable across additional fraud typologies.
KEY RESPONSIBILITIES
- Lead the graph analytics and fraud model enhancement workstream.
- Design and implement the graph feature engine using NetworkX, Neo4j or equivalent technologies.
- Build entity graphs across accounts, agents, devices, beneficiaries, merchants and geographic data points.
- Develop network features including centrality, community detection, multi-hop paths, fan-in, fan-out, off-ramp concentration and repeat-offender linkages.
- Detect fraud patterns such as aggregator or funnel activity, reward or enticement loops, layering, rapid fund movement, collusion, coordinated rings and suspicious exit points.
- Build and enhance Python machine learning pipelines for feature engineering, training, validation and scoring.
- Integrate graph, behavioural and historical interaction features into a point-in-time-correct feature vector for training and production scoring.
- Ensure graph metrics can be produced incrementally and through scheduled full recomputation.
- Review the existing fraud modus operandi and scheme library, identify data gaps and incorporate emerging fraud patterns.
- Improve model precision, recall and F1 performance while materially reducing false positives.
- Support model selection, threshold optimisation, champion-model promotion, drift detection and drift-triggered retraining.
- Ensure model outputs include explainable risk indicators, graph evidence, version metadata and complete audit lineage.
- Integrate investigator outcomes and labels into the feedback and retraining process.
- Work closely with the Data Scientist, MLOps Engineer, Solution / Data Architect and fraud investigation stakeholders.
- Provide technical leadership, review work completed by other team members and contribute to knowledge transfer and handover.
ESSENTIAL EXPERIENCE AND TECHNICAL REQUIREMENTS
- Proven hands-on experience in graph analytics using NetworkX, Neo4j or equivalent graph technologies.
- Advanced Python and machine learning pipeline development experience.
- A strong track record in fraud detection within fintech, banking, payments or regulated financial services.
- In-depth knowledge of fintech fraud typologies, social engineering fraud and financial crime patterns.
- Experience with graph algorithms, entity resolution, community detection and network-pattern analysis.
- Strong feature-engineering experience using transactional, behavioural and historical interaction data.
- Experience working with large, complex and sensitive financial datasets.
- Knowledge of supervised and unsupervised fraud modelling approaches.
- Experience evaluating models using precision, recall, F1 score, false-positive rate and related performance measures.
- Understanding of feature stores, point-in-time correctness, train-and-serve consistency, model versioning and auditability.
- Working knowledge of MLOps best practices, production model deployment, monitoring and drift control.
- Ability to lead a technical workstream while remaining actively involved in design, coding and delivery.
ADVANTAGEOUS EXPERIENCE
- Experience with online feature serving and low-latency fraud decisioning.
- Experience integrating model outputs with fraud case-management or investigation platforms.
- Exposure to PostgreSQL, Redis, Kafka, Docker, Airflow or equivalent technologies.
- Experience developing reusable fraud frameworks across multiple markets, business units or fraud typologies.
WAYS OF WORKING
- Work within an Agile technology team focused on continuous delivery.
- Participate in sprint planning, daily stand-ups, sprint reviews, retrospectives and demonstrations.
- Take ownership of assigned deliverables and communicate risks, dependencies and technical decisions clearly.
- Collaborate across data science, engineering, architecture, fraud and investigation teams.
WHO SHOULD APPLY?
- This role is suited to a senior graph data science specialist who has personally built graph-based fraud features and production machine learning solutions. General data science experience without hands-on graph analytics and proven fraud detection experience will not be sufficient.
- Apply through CapCircle. Your CV must clearly show your graph technology experience, Python and ML expertise, fraud detection projects, technical leadership and personal contribution to production solutions.
ADDITIONAL MANDATORY REQUIREMENTS
- A minimum of 5–8 years’ relevant professional experience is required.
- Candidates must hold a relevant tertiary degree and applicable professional or technical certifications related to the position.
- These are senior and advanced-level positions and are not suitable for interns, recent graduates or students.
- Candidates must be able to work on a hybrid basis from the office in Roodepoort, Gauteng, South Africa.
- Office attendance is required three days per week and is non-negotiable.
- Please only apply if you meet the required experience, qualification and hybrid working requirements.