Job: AI Intern

Company : Datafin; Johannesburg; On-site; ZAR 78,000 - 112,000; Full time

Job summary

Datafin in Johannesburg, Gauteng, South Africa, offers an applied-AI internship focused on preparing data for AI workloads and supporting a governed AI platform. Trainees work under supervision with a mentored pathway, learning how AI is used safely on sensitive data in a regulated business.

You will contribute to data preparation, testing AI use cases, and documenting validation evidence while adhering to data-minimisation rules and governance guidelines.

Qualifications

• Currently pursuing or recently completed studies in data science, CS, information systems, statistics, maths, engineering, or related fields.
• Exposure to APIs, notebooks, retrieval/vector search concepts, or LLM tooling.
• Evidence of an AI project solving a real problem and awareness of responsible-AI concerns.

Responsibilities

• Prepare, clean, validate, and document datasets used by AI and search workloads.
• Support the in-house governed AI platform with ingestion, indexing, configuration, and health checks.
• Test and evaluate AI-assisted use cases against defined success criteria.

Skills

Data Science

Python

APIs

Notebooks

Vector search

Education

Relevant degree field (in progress or completed)

Tools

LLM tooling

Job description

JOIN an applied-AI work on a structured, mentored pathway, rotating through the data engineering that feeds it as the next AI Engineer wanted by a provider of tailored Financial Solutions. You will prepare trusted data for AI workloads, help run an in-house governed AI platform, and test AI-assisted use cases under supervision and strictly inside policy. This is an engineering-and-governance internship , not a research post, and you will not be training bespoke models. What you will learn is how AI gets used safely and usefully on real, sensitive data in a regulated business. Interns who grow well here are first in line as the team expands, and the combined Data & AI

Responsibilities

• Prepare, clean, validate, and document the datasets that AI and search workloads consume.
• Support the in-house governed AI platform: content ingestion, indexing, configuration, and routine health checks under supervision.
• Test and evaluate AI-assisted use cases such as summarisation, classification, search, extraction, and automation against defined success criteria.
• Write small Python scripts and notebooks to automate repetitive data preparation and evaluation tasks.
• Record evaluation evidence: what was tested, what the AI produced, what a human reviewer changed, and why.
• Apply data-minimisation and anonymisation rules without exception, which means no personal information, credentials, or production data goes into unapproved tools.
• Contribute to documentation: use-case write-ups, prompt and configuration notes, data dictionaries, and testing evidence.

Qualifications

• Qualification in progress or recently completed in Data Science, Computer Science, Information Systems, Statistics, Mathematics, Engineering, or related.
• Exposure to APIs, notebooks, retrieval / vector search concepts, or any LLM tooling.
• Any project where you used AI to solve a real problem, especially if you can explain where it got things wrong.
• Portfolio, coursework, hackathon, or capstone evidence of data or AI problem-solving. Show us anything you have built.
• Basic awareness of responsible-AI concerns: bias, hallucination, privacy, and traceability.

Hiring organization

Datafin

Employment Type: internship

Job Location: Johannesburg,Gauteng,South Africa