Some new post ideas

It's been two and a half years since I started my first job as a data scientist, and I find that I have a few things to say by now. So I'm starting up the blog again!

Here are some post ideas, in no particular order:

• What I discuss with data science applicants in interviews
• What can go wrong if you don't use pipelines in model training
• The relationship between model complexity and validation robustness
• The changing role of data science in business
• Exotic types of data leakage
• Why neural nets are usually not very good
• Very general approaches to new data science projects
• Ethics of data and data science
• How to succeed by making yourself obsolete

We'll see what I can get to in my free time. See you soon.

- b