Data Science Career

by Chee Yee Lim


Posted on 2021-04-03



Collection of notes on data science career topics - everything on how to successfully manage oneself's career in data science field.


Managing Data Science Career

How to Succeed as a Data Scientist

  1. Have realistic expectations
    • Mindful of project budget, scope and time line.
      • It is usually more feasible to modify project scope to fit within the time line, than to modify project time line to fit project scope.
    • Most companies may not have the IT and/or data infrastructure to do data science projects.
      • Most of the time may be spent on setting the IT and data infrastructure up (and getting budget approved), rather than on building advanced models.
    • Most senior stakeholders are happy to see simple analytics charts/dashboards, rather than complex technical results that are unrelated to business (e.g. model accuracy).
      • Focus on first building dashboards/prototypes that show business values, before trying to build a complete solution (e.g. full stack production infrastructure) for a problem.
  2. Understand business values
    • Always try to understand the business context and potential business values that a data project can bring.
      • Remember everything in a business is a cost (including us data scientists). A cost in a business is only justifiable if it can either improve the revenue (e.g. bring in business) or reduce cost (e.g. improve business efficiency).
  3. Learn to navigate office politics
    • Build a relationship with key decision-makers/stakeholders in the company.
      • This does not mean brown-nosing/sucking up to higher ups.
      • This is to ensure a friendly relationship exists which enables healthy communications to potentially enable the higher ups to resolve issues.
      • Senior stakeholders may not understand the needs and/or what problems can data science solves. Having a good relationship with them may help them understand data science better.
      • Most senior stakeholders do not care about the technical aspects of data science. It is easier to win favour with them by doing mundane or basic tasks such as automation and analytics reporting.
    • Always try to first improve the situation by reaching out to key people.
      • Sometimes a change may not happen. In this case, staying low to ignore the problem or exiting the job may be the solutions.

Data Science Interviews

Questions to ask Interviewers

  1. Organisation data strategy
    1. Who is bought into the data science initiative at the most senior level of the organisation?
    2. How long have the organisation been working with data science? And what are the major achievements?
  2. Team structure
    1. How many people are in the data team?
    2. Are there supports provided by data engineers/analysts/engagement managers/DevOps engineers, or are the data scientists expected to do all these roles?
  3. Project structure
    1. How are the data projects structured? In terms of types of people involved, time lines and deliverables.
  4. IT/Data infrastructure
    1. What IT and data infrastructures are available and used regularly for projects? (Databases, servers, clusters and cloud)
    2. How well is the integration between all databases / data sources? Isolated system or well integrated (e.g. into data warehouse).

Sources