Data Scientist - Sandton at Discovery Limited

Responsibilities include:

• Working with huge quantities of unstructured text data from a variety of sources.
• Being responsible for and owning, along with the rest of the team, the delivery of projects from inception through to business adoption, including:
• Completing reviews of relevant academic literature and industry releases.
• Prototyping code for data science and ML systems, particularly those using NLP and LLMs.
• Running scientific experiments to evaluate different models and system components within prototypes during development.
• Working with data engineers and systems architects to deploy solutions should the experiments prove successful.
• Working with data engineers and systems architects to monitor deployments and maintain models and systems.
• Presenting analyses and project updates to both technical and business audiences.
• Keeping an open mind and looking for new opportunities for the use of existing datasets and tools, as well as new ones, for novel business applications.
• Supporting existing deployments of NLP models, including sentiment and topic models.
• Constantly reviewing academic literature and industry releases to ensure that our research and systems remain state-of-the-art.

Personal Attributes  

• A creative and eager attitude to unearthing valuable insights and generating value for Discovery clients.
• Enthusiasm for building solutions and analyses scientifically and robustly.
• Ability to balance multiple priorities and step back to see how your work fits into the wider business context.
• Aligned to Discovery values and core purpose.

Technical Skills

Required

• SQL and working with databases.
• Python for data science and machine learning.
• Ability to formulate a clear problem statement, develop a plan for tackling it, and clearly communicate findings verbally, visually, and in writing.

Advantageous

• Version control (Git).
• Experience with R.
• Experience with using and/or developing NLP packages and models.
• Experience with TensorFlow and/or PyTorch.
• Experience with using and/or training LLMs.
• Experience with Spark and/or Dask.

Education and Experience

• Honors or Master’s degree in Computer Science, Mathematics, Statistics, Data Science, Actuarial Science, Statistics, Operations Research, Industrial engineering, Applied Mathematics, or similar quantitative field. A PhD degree would be advantageous. Other qualifications will also be considered if accompanied by relevant experience.
• We will consider candidates at all levels of experience.