Data Scientist at Discovery Limited

Key Purpose

• To analyse large amounts of data to provide core business value and key insights and enhance data-driven business processes and decisions.

Areas of responsibility may include but not limited to

• Building models and developing algorithms to support new key initiatives, eg. better detect and understand driving actions, driver intent, forecasting driver behavior
• Maintaining and improving driver scoring and analytics framework within AWS
• Data mining and cleaning
• Interpret and disseminate information via reports
• Researching and prototyping new technologies and methods that are relevant to the team’s initiatives
• Analysis, presentation and communication of findings and data in a way that provides actionable insights and can be understood by larger parts of the business
• Providing business stake holders with relevant telematics data in a useable format
• Comfortable working in Linux or MacOS environment
• Write and test high quality Python code that will run in a production environment.  Familiarity with standard Python libraries

Skills

• Programming experience in Python (preferred)
• Visualisation tools and packages (eg. Seaborn, dash, plot.ly)
• Experience with the Python data science stack is required (numpy, pandas, jupyter, scipy, etc)
• Working knowledge of SQL
• Comfortable working with Unix-based operating systems (especially working from the terminal)
• Experience with AWS is advantageous
• GIS experience is advantageous

Knowledge

• Amazon Web Services – adv
• Familiarity with machine learning packages i.e. Tensor flow, Keras, FastAI
• Understanding of statistics and familiarity with statistical computing packages (eg. Scipy, numpy)
• Experience in data mining, feature extraction
• Modern machine learning techniques, in particular working with time series data, geospatial data, device data and natural language processing
• Algorithm analysis and code optimization
• Understanding of techniques required to work with extremely large datasets (larger than RAM)
• Ability to work with both structured and unstructured data, and ability to combine data from diverse sources (APIs, databases, web-scraping, for eg.)

Education and Experience

• Matric
• Masters or Hons in applied mathematics, engineering, physics, or computer science
• Recent graduate (Honors or above)
• Strong statistics and analytical skills
• Prior programming experience required
• Working and/or research experience is preferable, but not required