Data Consultants: Fraud & Forensics – Four Contract Opportunities
CapCircle is recruiting four highly experienced data professionals for a large-scale fintech fraud detection and model modernisation programme.
These are senior and advanced-level positions for specialists who can work hands-on, take ownership of critical delivery areas and operate effectively in a fast-paced, regulated financial environment.
Positions available: Four
• Senior Graph Data Scientist / ML Lead
• Solution / Data Architect
• MLOps Engineer
• Data Scientist
Your CV must clearly show your relevant fintech or financial-services fraud experience, technical stack, seniority and hands-on contribution to similar production projects.
THE PROGRAMME
The team will enhance an existing social engineering fraud detection solution into a high-precision, end-to-end capability covering data ingestion, feature engineering, model scoring, inline decisioning, case management and the investigator feedback loop.
The objective is to materially reduce false positives, improve recall and create a scalable, reusable platform that can support additional fraud typologies without redesigning the core architecture.
The work is divided into two connected areas:
- Multi-model server architecture: Design and implement a production-grade environment for multiple fraud models, data stores and supporting services, with orchestration, resource isolation, containerisation, observability and scalability.
- Fraud model enhancement: Improve fraud detection through behavioural features, historical interaction data, graph analytics, anomaly detection, model selection and drift control.
AVAILABLE POSITIONS
- SENIOR GRAPH DATA SCIENTIST / ML LEAD
This person will lead the fraud model enhancement workstream and own the graph feature engine and model improvements.
Key responsibilities:
- Design and implement graph-based fraud detection features using NetworkX, Neo4j or equivalent graph technologies.
- Build entity graphs across accounts, agents, devices, beneficiaries, merchants and locations.
- Develop network metrics such as centrality, communities, multi-hop paths, fan-in and fan-out, off-ramp concentration and repeat-offender linkages.
- Identify aggregator, funnel, reward-loop, layering, rapid-movement, collusion, community and exit-point patterns.
- Build and enhance Python machine learning pipelines.
- Integrate graph, behavioural and historical interaction features into a single, point-in-time-correct feature vector.
- Improve model precision and recall while reducing false positives.
- Ensure graph features are explainable, versioned, auditable and available to fraud investigators.
- Support model selection, performance monitoring, drift detection and retraining.
Required experience:
- Advanced hands-on graph analytics experience using NetworkX, Neo4j or equivalent.
- Strong Python and machine learning pipeline development experience.
- Proven fraud detection experience within fintech, banking, payments or financial services.
- Experience with graph feature engineering, entity resolution and network-pattern detection.
- Ability to lead a technical workstream while remaining hands-on.
- SOLUTION / DATA ARCHITECT
This person will lead the architecture workstream and own the target-state platform and data design.
Key responsibilities:
- Design a multi-model-per-server architecture that supports fraud detection, scoring, feature engineering, graph processing, model selection and case management.
- Define a two-instance PostgreSQL topology separating detection and investigation workloads from offline feature engineering and retraining.
- Design the offline and online feature stores, low-latency serving path and inline decisioning capability.
- Establish resource isolation policies for CPU, memory and I/O to support predictable concurrent performance.
- Define availability, latency, failover, degradation, retention, partitioning and data-reconstructibility requirements.
- Design for train-and-serve parity, data lineage, auditability and point-in-time correctness.
- Produce the scalability roadmap and future cloud migration approach.
- Ensure the architecture is reusable for additional fraud models and typologies.
- Guide concurrency, load, latency, replay, backlog recovery and failover testing.
Required experience:
- Proven experience architecting production data and machine learning platforms in a regulated financial environment.
- Strong knowledge of PostgreSQL architecture, feature stores, distributed data processing and low-latency decisioning.
- Experience designing multi-model infrastructure, resource isolation and scalable data platforms.
- Understanding of data governance, security, audit, privacy and regulatory requirements.
- Experience developing cloud migration and platform scalability roadmaps.
- MLOPS ENGINEER
This person will build and optimise the infrastructure required to deploy, orchestrate, serve and monitor multiple fraud models
Key responsibilities:
- Implement multi-model infrastructure with isolated and containerised model pipelines.
- Containerise model components using Docker and manage dependencies across environments.
- Configure Airflow or an equivalent orchestration framework for scheduling, sequencing, retries, monitoring and failure isolation.
- Build and optimise streaming or near-real-time data pipelines using Kafka or equivalent technologies.
- Implement Redis or an equivalent online store for low-latency feature serving and inline decisioning.
- Optimise PostgreSQL performance, indexing, partitioning, retention and maintenance.
- Monitor model-serving latency, cycle duration, lag, backlog, throughput and platform availability.
- Support automated deployment, model versioning, observability, drift-triggered retraining and rollback controls.
- Conduct concurrency, load, stress, replay and failover testing.
- Support graph data pipelines and incremental and full graph recomputation.
Required experience:
- Strong MLOps and production machine learning infrastructure experience.
- Advanced experience with Docker and Airflow or equivalent orchestration technologies.
- Experience with Kafka or similar streaming platforms and Redis or equivalent low-latency stores.
- Strong database optimisation and PostgreSQL experience.
- Experience deploying low-latency model-serving and near-real-time decisioning solutions.
- Working knowledge of graph data engineering.
- Understanding of model monitoring, drift control, auditability and data lineage.
- DATA SCIENTIST
This person will support the fraud model enhancement workstream, focusing on behavioural and historical interaction features, anomaly detection, model selection and performance benchmarking.
Key responsibilities:
- Review and refine the existing fraud modus operandi and scheme library.
- Engineer behavioural, transactional and historical interaction features.
- Analyse prior transaction and communication patterns to distinguish genuine relationships from suspicious activity.
- Develop and assess unsupervised anomaly-detection approaches for transactional data.
- Support entity linking, explainable risk scoring and fraud pattern detection.
- Evaluate models and thresholds using precision, recall, F1 score and false-positive-rate measures.
- Conduct before-and-after performance benchmarking.
- Support model selection, champion-model promotion, drift monitoring and drift-triggered retraining.
- Implement deduplication and data-integrity controls.
- Support investigator feedback integration and end-to-end output validation.
Required experience:
- Strong data science experience within fraud, financial crime, banking, payments or fintech.
- Proven experience with unsupervised anomaly detection on transactional data.
- Strong Python, feature engineering, model evaluation and statistical analysis skills.
- Experience working with large transactional datasets and behavioural patterns.
- Knowledge of graph analytics or graph-derived features is advantageous.
- Experience with model monitoring, drift control and performance benchmarking.
REQUIREMENTS FOR ALL FOUR POSITIONS
Candidates must have:
- Strong knowledge of fintech fraud typologies, social engineering fraud and financial crime patterns.
- A sound understanding of MLOps best practices and production machine learning environments.
- Experience working with sensitive financial and transactional data in a regulated environment.
- The ability to translate complex technical requirements into scalable, practical solutions.
- Strong problem-solving, documentation and stakeholder-management skills.
- Experience working in Agile delivery teams using sprint planning, daily stand-ups, reviews, retrospectives and demonstrations.
- A mature, delivery-focused approach with the ability to take ownership and collaborate across data science, engineering, architecture, fraud and investigation teams.
- Strong communication skills and the ability to explain technical decisions and model outcomes clearly.