AI Engineer

Job Title: AI Engineer (Applied AI and RAG Systems)

Location: Woodmead, Sandton, South Africa

Primary Purpose (Role):

We are seeking a talented AI Engineer to join our dynamic team to develop and implement applied AI solutions within a collaborative, team‑based delivery model. This role focuses on building, integrating, and supporting AI components while ensuring solutions are well documented, reproducible, and transferable to fellow AI and software engineers.

You will contribute to digitalisation and innovation initiatives, working closely with your team to deliver shared AI solutions. The role emphasises growth in applied AI capabilities through collaboration, pairing, and hands‑on delivery. Architectural decisions are made jointly within the Digital Factory and IT Architecture teams, as the role does not own AI solutions in isolation.

• RAG Architecture and Design
  • Implement and contribute to the evolution of end‑to‑end RAG architectures, including ingestion, indexing, retrieval, reranking, and generation layers.
  • Understand and manage trade‑offs between latency, accuracy, cost, and maintainability.
• Ingestion and Indexing
  • Develop ingestion workers for structured and unstructured data (PDFs, Office documents, HTML, APIs, and databases).
  • Implement chunking, metadata enrichment, embedding generation, and versioned indexing.
  • Manage re‑ingestion, incremental updates, and data freshness.
• Retrieval and Query Pipelines
  • Create query workers performing retrieval, filtering, reranking, and context assembly.
  • Implement hybrid search (vector and keyword), metadata filtering, and scoring strategies.
  • Integrate large language models (LLMs) for reasoning, summarisation, and tool use.
• LLM Integration
  • Integrate and evaluate commercial and open‑source LLMs (e.g. OpenAI).
  • Develop prompt templates, function/tool calling, and structured outputs.
  • Apply guardrails, grounding, and citation strategies to ensure reliability.
• Infrastructure and Operations
  • Build and utilise scalable services using queues, workers, and asynchronous processing.
  • Work with vector databases (e.g. Pinecone, Milvus).
  • Implement logging, monitoring, evaluation, and cost controls for LLM workloads.
• Quality and Evaluation
  • Conduct evaluation strategies to measure retrieval quality and answer accuracy.
  • Debug hallucinations, retrieval failures, and latency bottlenecks.
  • Continuously improve system accuracy and reliability.
• Engage with stakeholders to understand business needs and translate these into technical requirements.
• Adhere to best practises in code development, documentation, and deployment across environments.
• Collaborate with product owners and stakeholders to align AI‑enabled solutions with business processes.
• Contribute across the full systems lifecycle, including design, coding, testing, implementation, maintenance, and support of software delivered on time and within budget.
• Follow Agile methodologies to deliver fit‑for‑purpose solutions.
• Knowledge Sharing and Capability Enablement
  • Ensure all AI systems are documented, reproducible, and understandable by other engineers.
  • Participate in regular technical walkthroughs and demonstrations for team members.
  • Actively share knowledge and learn new skills through pairing, reviews, and written artefacts.

• Strong software engineering background, with Python expertise essential.
• Experience building production APIs and background workers.
• Hands‑on experience with RAG systems beyond prototype stage.
• Understanding of embeddings, vector search, and retrieval strategies.
• Experience with chunking, reranking, and context‑window optimisation.
• Familiarity with LLM orchestration frameworks (e.g. Llama Index) or equivalent custom implementations.
• Experience with vector and traditional databases.
• Knowledge of message queues and asynchronous processing (e.g. Celery).
• Familiarity with Microsoft Azure Dev Ops for version control.
• Experience with document parsing and OCR pipelines.
• Adherence to coding standards and supporting their adoption across the team.
• Ability to identify and raise technical and delivery risks, supporting mitigation efforts.
• Works under the guidance of senior engineers within established AI…