I think being a Data Scientist is pretty similar to my postdoc. During my postdoc I was working with data, building mathematical models, and doing lots of programming, which is exactly what I do now.
During my postdoc I programmed in MATLAB and Fortran, while I now use Python. Although it wasn’t difficult for me to transition because the programming languages are all quite similar, it was challenging convincing hiring managers of this. I wish I had spent more time with Python sooner so that I was more confident in the early interviews.
Presenting work is also important in both a postdoc and in industry, although the audience is different. In academia, conferences are usually filled with experts from the same field. However, the best talks I ever went to were the ones that had clear visualisations and presented interesting results, rather than overly complex mathematical methods.
These are the skills needed when presenting in industry. You need to be able to find the interesting and relevant results so that stakeholders are engaged as well as build confidence that the methodology is robust without going into too much technical detail.
Data Science projects are typically 3-6 months long and the success is measured on being able to deliver the results. It’s important to agree with stakeholders what deliverables are achievable in the timeframe at the outset of the project.
When planning how you’re going to achieve the objectives, it’s also important that you have an idea of what work you’re going to perform as well as making backup plans in case those methods don’t work. This is very similar to writing a grant proposal. So, although your grant might not be funded, you’re still learning valuable transferable skills.