Read voraciously. A membership in IEEE and ACM journals will keep your stack full although it can be a pain sorting through all of the wheat. Keep track of book

Read voraciously. A membership in IEEE and ACM journals will keep your stack full although it can be a pain sorting through all of the wheat. Keep track of books that interest you in subject areas that seem challenging and interesting: if you're interested in data that will definitely be in the realm of statistics, combinatorics, linear algebra. Do exercises. Take 30-40 mins a day to do practice problems. Use a site like codewars, leetcode, even past advent of code challenges. Do the exercises in the text books you read. Write. Write about what you learn. Use your writing to explain it to someone who is encountering the subject area for the first time. Doesn't matter what you write it on: a paper notebook is just as serviceable as a blog. The act of reinforcing what you discover by writing it down helps to reinforce it and challenges your knowledge by explaining it to someone else. Whatever you do make sure that the act of writing works for you: it must be seamless and flow easily. Don't over-engineer the perfect writing system that use a ton of frameworks and require maintaining deployment scripts. You will avoid writing if it is hard to do. Writing itself is hard enough and that's what you need to focus on. If you want to go the blog route make sure it's no harder than opening up a text file and running a simple script when you're done. Lastly, find a mentor. It can be someone outside of work -- and in your case sounds like it will have to be. Contribute to an open source project you use regularly. Go to meetups in your local area or online. Find someone you can talk shop with. And when the opportunity arises ask if they'd be interested in meeting up once in a while to review your work or talk about what you've discovered.