SENIOR MLOPS ENGINEER – FRAUD & FORENSICS
Level: Senior / Advanced
- CapCircle is recruiting a highly skilled MLOps Engineer for a fintech fraud detection and model modernisation programme.
- This is a hands-on engineering role for someone experienced in multi-model infrastructure, containerisation, orchestration, database optimisation, low-latency model serving and streaming or near-real-time pipelines. Graph data engineering knowledge is required.
ROLE OBJECTIVE
Build, deploy and optimise the production infrastructure required to run multiple fraud models and supporting services reliably on shared infrastructure. The platform must sustain recurring detection cycles and low-latency decisioning while isolating scoring, feature engineering, graph computation and retraining workloads.
KEY RESPONSIBILITIES
- Implement multi-model infrastructure for fraud detection, feature engineering, scoring, retraining and case-management services.
- Containerise model pipelines and supporting components using Docker, with isolated environments and controlled dependency management.
- Configure Airflow or an equivalent orchestration framework for scheduling, sequencing, retry logic, monitoring and failure isolation.
- Build and maintain reliable batch, streaming and near-real-time data pipelines.
- Implement Kafka or an equivalent streaming technology for event-driven or near-real-time processing.
- Implement and optimise Redis or an equivalent online store for low-latency feature serving and inline decisioning.
- Support PostgreSQL architecture across transaction, case-management and offline feature-store workloads.
- Optimise database queries, indexes, partitions, autovacuum, maintenance routines and storage performance.
- Implement incremental and idempotent data-loading processes, including controls for late-arriving records and replay.
- Operationalise behavioural, historical interaction and graph features for both model training and production scoring.
- Support incremental graph feature computation and scheduled full graph recomputation.
- Ensure consistency between offline and online features and prevent training-serving skew.
- Deploy and monitor low-latency model-serving services with defined latency, availability, failover and degradation controls.
- Implement observability for cycle duration, lag, backlog, throughput, failures, resource consumption and serving latency.
- Establish CPU, memory and I/O controls to prevent contention between detection, backfill and retraining workloads.
- Support model versioning, controlled deployment, champion-model promotion, rollback and drift-triggered retraining.
- Implement audit, lineage and decision-logging requirements for model outputs.
- Perform concurrency, peak-load, sustained-cycle, latency, backlog-recovery, replay and failover testing.
- Document deployment processes and contribute to scalability, cloud migration and production handover.
- Work closely with the Solution / Data Architect, Graph Data Scientist / ML Lead, Data Scientist and fraud stakeholders.
ESSENTIAL EXPERIENCE AND TECHNICAL REQUIREMENTS
- Proven hands-on MLOps or production ML engineering experience.
- Advanced Docker experience, including containerisation of model pipelines and dependency isolation.
- Strong Airflow experience or equivalent workflow orchestration expertise.
- Experience building streaming or near-real-time pipelines using Kafka or equivalent technology.
- Experience implementing Redis or equivalent low-latency online stores.
- Strong PostgreSQL administration, query optimisation, indexing, partitioning and performance-tuning skills.
- Experience deploying and operating low-latency model-serving or decisioning solutions.
- Experience supporting multiple models and workloads on shared infrastructure.
- Strong understanding of CI/CD, environment management, monitoring, logging, alerting and production support.
- Working knowledge of graph data engineering and graph-processing pipelines.
- Understanding of feature stores, point-in-time correctness, train-and-serve consistency and data lineage.
- Knowledge of model monitoring, versioning, drift control, retraining and rollback practices.
- Strong knowledge of MLOps best practices and fintech fraud typologies.
- Experience working with sensitive data and production platforms in regulated financial environments.
ADVANTAGEOUS EXPERIENCE
- 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 production reliability, engineering quality and assigned deliverables.
- Communicate technical risks, dependencies and performance constraints clearly.
- Collaborate across architecture, data science, data engineering, fraud and investigation teams.
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 production reliability, engineering quality and assigned deliverables.
- Communicate technical risks, dependencies and performance constraints clearly.
- Collaborate across architecture, data science, data engineering, fraud and investigation teams.
WHO SHOULD APPLY?
- This role is for an experienced, hands-on MLOps Engineer who has deployed and supported production machine learning systems. Candidates whose experience is limited to data science notebooks, model development or general DevOps without production ML infrastructure will not meet the core requirements.
- Apply through CapCircle. Your CV must clearly show your production MLOps projects, technical stack, scale, performance requirements and personal contribution to implementation and support.
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.