Cape Town City Centre: Junior-Mid Agentic AI Engineer (Python) (Remote)

About the Role

We are seeking a Junior-Mid Agentic AI Engineer (Python) to join our team in Cape Town City Centre. As a key member of our engineering team, you will design, develop, and deploy agentic AI systems and LLM-powered applications in production environments. You will work closely with the ML Engineering team on model evaluation, testing, and continuous improvement, building and optimizing RAG pipelines while contributing to the design of agentic architectures, tool-use patterns, and orchestration frameworks.

Key Responsibilities

• Design, develop, and deploy agentic AI systems and LLM-powered applications in production environments.
• Build and optimize RAG (Retrieval-Augmented Generation) pipelines, including document ingestion, chunking strategies, embedding models, and retrieval mechanisms.
• Integrate and manage vector databases (e.g., Pinecone, Weaviate, Qdrant, Milvus, ChromaDB) for efficient similarity search and knowledge retrieval.
• Develop and maintain Backend services and APIs (primarily in Python) to serve AI models and agent workflows.
• Work with the ML Engineering team on model evaluation, testing, and continuous improvement.
• Contribute to the design of agentic architectures, tool-use patterns, and orchestration frameworks.
• Implement guardrails, monitoring, and observability for LLM-based systems in production.
• Collaborate on MLOps practices including model registry, experiment tracking, and CI/CD for ML pipelines.

Requirements

• 2–3 Years of professional experience in AI/ML Engineering or a closely related role.
• At least one production-level Agentic project — you’ve built, deployed, and maintained an agent-based system that serves real users or real workloads.
• Solid foundation in general Machine Learning — supervised/unsupervised learning, model training, evaluation metrics, and data preprocessing.
• Hands-on experience with LLM application development — prompt engineering, fine-tuning, function/tool calling, and structured output generation.
• Working knowledge of the agentic stack — agent frameworks, tool integration, memory management, planning and reasoning patterns, and multi-step orchestration.
• Practical experience with RAG architecture — end-to-end pipeline design, embedding models, retrieval strategies, and re-ranking.
• Exposure to vector databases — setup, indexing, querying, and performance tuning.
• Strong Python skills — clean, well-structured, production-quality code.
• Comfortable with async programming, REST APIs, and standard data/ML libraries.

Qualifications

• Formal education/certifications not specified.