Description:
We are seeking an experienced AI/ML engineer to join our dynamic merchant payments team in Stellenbosch/Cape Town. This is a hybrid role requiring in-office presence three days per week.
As part of a collaborative and high-performing team, you will design, build, and productionise secure, scalable AI/ML solutions with a strong focus on merchant payments and financial transaction processing. Working at the intersection of data engineering and machine learning, you’ll contribute to projects that demand innovation, compliance, and technical excellence.
If you thrive on solving complex challenges, enjoy working alongside motivated professionals, and want to deliver AI systems that make a real impact, this role is for you.
Key responsibilities and output areas include:
LLM Solution Engineering
- Architect and implement RAG pipelines over payments and merchant data (tickets, chargebacks, settlement statements, documentation).
- Fine-tune and evaluate domain-adapted models (prompting, adapters/LoRA, retrieval, re-ranking, and guardrails).
- Deploy secure, observable model services behind internal APIs with SLA/SLOs and A/B or canary rollout.
Data Engineering for AI
- Build events drive pipelines (Kafka) and ELT into feature/embedding stores; manage data contracts, schema evolution, and quality gates end to end.
- Optimise storage/layouts and retrieval for vector indexes and low-latency use cases.
MLOps & Reliability
- Implement training/inference CI/CD, model/version registries, dataset lineage, drift/guardrail monitoring, red teaming, and safety filters.
- Define offline and online evaluation (factuality, toxicity, PII safety, and business KPIs such as deflection, turnaround time, and fraud catch uplift).
Security & Compliance
- Engineer solutions consistent with PCI DSS expectations when handling card data (segmentation, least privilege, logging/monitoring, encryption, and vulnerability management).
- Partner with security, risk, and compliance teams for design reviews.
Collaboration & Delivery
- Partner with Merchant Ops, Risk/Fraud, Disputes, Engineering, and Legal teams to translate problem statements into shipped AI features.
- Contribute to AI governance (model cards, usage policies, evaluation packs).
Key Requirements
- 6+ years in data or ML engineering, with at least 2+ years hands on in LLMs / modern NLP (RAG, vector search, embeddings, prompt engineering, fine-tuning).
- Proven experience in production deployments of LLM-backed services (latency, cost, monitoring, rollback).
- Technical Skills
- Strong proficiency in Python (data + serving) and SQL.
- Experience with event streaming (Kafka) and workflow orchestration (Airflow/Dagster).
- Cloud expertise (AWS): containerised workloads, infrastructure as code (Terraform), and Kubernetes (Rancher-managed environments acceptable).
- Experience building and operating data pipelines (batch + streaming), dimensional models, and feature/embedding stores.
Domain Knowledge
- Ability to design solutions aligned with PCI DSS controls (or willingness to upskill rapidly).
- Payments/merchant domain exposure (chargebacks, disputes, settlements, reconciliation, 3DS, fraud signals) is advantageous.
- Soft Skills
- Strong problem-solving and analytical ability.
- Ability to work effectively in agile teams.
- Clear communication with technical and non-technical stakeholders.
Education
- Bachelor’s degree in computer science, information systems, or a related field.
- Relevant certifications in AI/ML, cloud, or payment systems are advantageous.
Nice to Have
- Experience with vector databases (e.g., OpenSearch k NN, pgvector, Pinecone), rerankers, safety/guardrail frameworks.
- Observability for AI: prompt/version telemetry, evaluation pipelines, and human-in-the-loop feedback.
- FinOps mindset: model/runtime cost optimisation at scale on AWS.
- Exposure to performance tuning, observability tools, or A/B testing.
Please Note:
As all iqbusiness roles require honesty in the handling of or access to cash, finances, financial systems, or confidential information, our recruitment process requires that the following background checks be completed: credit, criminal, ID, and qualification verification.
iqbusiness is committed to sustainable growth and transformation; we embrace diversity and employ previously disadvantaged individuals.
Requirements:
- Architect and implement RAG pipelines over payments and merchant data (tickets, chargebacks, settlement statements, documentation).
- Fine-tune and evaluate domain-adapted models (prompting, adapters/LoRA, retrieval, re-ranking, and guardrails).
- Deploy secure, observable model services behind internal APIs with SLA/SLOs and A/B or canary rollout.
- Build events drive pipelines (Kafka) and ELT into feature/embedding stores; manage data contracts, schema evolution, and quality gates end to end.
- Optimise storage/layouts and retrieval for vector indexes and low-latency use cases.
- Implement training/inference CI/CD, model/version registries, dataset lineage, drift/guardrail monitoring, red teaming, and safety filters.
- Define offline and online evaluation (factuality, toxicity, PII safety, and business KPIs such as deflection, turnaround time, and fraud catch uplift).
- Engineer solutions consistent with PCI DSS expectations when handling card data (segmentation, least privilege, logging/monitoring, encryption, and vulnerability management).
- Partner with security, risk, and compliance teams for design reviews.
- Partner with Merchant Ops, Risk/Fraud, Disputes, Engineering, and Legal teams to translate problem statements into shipped AI features.
- Contribute to AI governance (model cards, usage policies, evaluation packs).
- 6+ years in data or ML engineering, with at least 2+ years hands on in LLMs / modern NLP (RAG, vector search, embeddings, prompt engineering, fine-tuning).
- Proven experience in production deployments of LLM-backed services (latency, cost, monitoring, rollback).
- Technical Skills
- Strong proficiency in Python (data + serving) and SQL.
- Experience with event streaming (Kafka) and workflow orchestration (Airflow/Dagster).
- Cloud expertise (AWS): containerised workloads, infrastructure as code (Terraform), and Kubernetes (Rancher-managed environments acceptable).
- Experience building and operating data pipelines (batch + streaming), dimensional models, and feature/embedding stores.
- Ability to design solutions aligned with PCI DSS controls (or willingness to upskill rapidly).
- Payments/merchant domain exposure (chargebacks, disputes, settlements, reconciliation, 3DS, fraud signals) is advantageous.
- Soft Skills
- Strong problem-solving and analytical ability.
- Ability to work effectively in agile teams.
- Clear communication with technical and non-technical stakeholders.
- Bachelor’s degree in computer science, information systems, or a related field.
- Relevant certifications in AI/ML, cloud, or payment systems are advantageous.
- Experience with vector databases (e.g., OpenSearch k NN, pgvector, Pinecone), rerankers, safety/guardrail frameworks.
- Observability for AI: prompt/version telemetry, evaluation pipelines, and human-in-the-loop feedback.
- FinOps mindset: model/runtime cost optimisation at scale on AWS.
- Exposure to performance tuning, observability tools, or A/B testing.
24 Feb 2026;
from:
careers24.com