Begin learning machine learning
If you go to the competition page on Kaggle, you can find a number of open competitions. If you scroll down to the bottom, you can find ~4 competitions with a light blue (or green) tab, and a “101”. Those are tutorial Kaggle competitions that contains lots of codes that helps beginner start with machine learning and classification/regression. Each of them sort of address different type of problem. “Digit Recognizer” is a image classification problem. “Titanic Disaster” is a typical binary classification problem. “Facial Keypoint Detection” is also an image related problem. Beginners can follow the instructions to download data, load data, play around with them and try fitting some models as described in the tutorials.
slightly experienced users
For slightly experienced users, they can move further up the ladder of competitions and get exposed to more variety of problems. At this stage, it is still a learning period. But now, instead of learning from the tutorials provided, you can learn from other users’ scripts. You can see what kind of tools they are using, why they use them, how they preprocess the data, what kind of models did they use, how they tuned their parameters, what are the performances of the model, etc.
there are 3 major categories of competitions
Then, there are 3 major categories of competitions: Featured, Recruitment, and Research. Featured competitions usually have a very high winning awards. Those are usually problems that famous companies have and they are looking for solutions from data scientists all around the world. Recruitment competitions, as explained by the tab, is used by companies to recruit potential employees in machine learning and data science area. Those who did well in the competitions will have the chance to get interviews from host companies. Some of the companies include: Walmart, Facebook, etc. Research competitions usually lasts longer since the deadlines for researches are less important than the outcome of the research. The problems are usually harder, and they usually have a large impact in some of the research areas. For example, the right whale recognition one is hosted to find a better solution to keep track of the very smaller number of right whales that are currently existing on the earth. Users can participate in various of competitions to compete their algorithms with others, learn from others and hopefully get some awards :).
Data Science Jobs Board
The above are for regular users, data scientists, machine learners. Notice that Kaggle also has a job posting board: Data Science Jobs Board where recruiters who are looking for data scientists/ machine learners can post their jobs there and specifically target the right group of people.