This is a typically offered solution. There are many websites out there like leetcode, project euler, hacker rank, interviewcake and codeeval that offer what they usually refer to as coding interview prep questions, and if your main goal is to land a job, you will probably learn a lot of specifically useful things, as well as how to think of specific problems to find clever solutions. There are a lot of problems, and its unlikely you'll transcend what these sites have to offer for a long time. The focus on sticking to the core of your language, with no fancy gems or extraneous tools can really serve to ground you in reality and teach you the finer details of your core skills.
These benefits come with a few drawbacks. Its hard to showcase on a resume, and while some sites, such as hacker rank offer a score you can show, unless its truly impressive its probably not something an interviewer is going to care about excessively. It also doesn't really teach you what building big, full fleshed projects feels like, and leaves you stranded in the land of your specific language, rather than bringing in interesting frameworks, and new tools.
If you like puzzles though, Coding challenges will provide small bite sized doses that provide a tangible feeling of progress, and that's important.
Build for yourself
We all have it, the dream project we all have wanted to build, maybe its something that's been built and implemented often, like Conway's Game of Life, or a simple maze program. Or maybe its more ambitious, like a fully fledged simulation of the behavior of twitter users, or a method to get push notifications whenever someone posts on a forum you like. Whatever it is, take this project and break it down into its component parts. Figure out the details, break down what you need to learn how to do, and what you know what to do. Then build it. Start with the smallest parts, and learn what you don't know. An evening of messing around with a database can teach you more than reading a few thousand words on it, and you won't know how to properly build routes until someone comes along and breaks the website you so cleverly crafted. This will teach you many things, be it deployment, or database usage, as well as training you on the number one thing every programmer needs to know. How to learn. You will break things. Things will not work, but you will learn.
This comes with a few drawbacks as well. A lot of what you'll learn is not immediately applicable. You'll run into paywalls and insurmountable problems. Getting a product you're satisfied with takes time. Your github account may wind up littered with the corpses of half finished programs, perhaps discouraging interviewers who don't understand the walls and limitations you've hit. Your raspeberry pi will catch fire when you screw something up badly enough. Interviewers won't care what you've built if you can't figure out reversing a linked list.
Most people who learn to code do so out of a desire to build. If building is your thing, then *build*, you'll be the better for it.
Build for others
I once read on a stack overflow question, that the best way to learn was to go out there and find problems people needed solving, and to figure out how to solve them. Its taken me some time and a growth in my skills to realize that even a very junior coder has skills very few people have and that problems that seem very difficult can be almost trivial with the right tools. When you're done building all your side projects, or when they've grown burdensome, or if you simply don't have anything you really want to build for *you*, go out and find problems that need solving. Earlier this week, i discovered that a certain forum I like has no good working mobile app, and no push notifications. I set out to try to figure out solutions to these problems. Will I build a killer app? Probably not, but I might use twilio to replicate push notifications for interested users. This has many of the same benefits as building for yourself, with the added benefit of teaching you what interacting with real users and their nebulous requirements is actually like
This is not free of burdens. Some problems haven't been solved because they are very hard to solve, or because the owner of the product doesn't wish to fix the issues. Sometimes you'll run into the real issue of hosting and maintaining a product other people use, and problems of ethics. You may find that what you've built relies on others, and doesn't last, consigning your hard work to obscurity and uselessness
The bottom line, is that with some basic coding skills, you can tackle problems others find impossible. By doing so, you can learn more, and have actual deployed projects to show interviewers.
This is not an exclusive list by any means, just a resource I wish I'd had when I started googling around for answers to this question. Some or all of these methods may not work for you, they're simply what work for me. At the very least, you should be asking "what can I do?", and if you are, you're already on the right track.