Hello, I am a student of Data Science and this week I was offered my first job in the area. The job is in a consulting company to work with a partner that has bought the software DataRobot and they need a team to operate it.
I never heard about this software before. Did some research and apparently this software do the job of a data scientist. So I am afraid that I will be a monkey pressing buttons and not really work applying my data science skills. What do you guys think?
I would advise you to look at it as complementary to what you do, not as a threat to replace you. How many more experiments and iterations can you do based on your ideas? Particularly if you take advantage of using the API? With you and DataRobot together, how much more accurate can your models become? How many more can you deploy and integrate into real world use case pipelines, where their value is captured? I would argue you are at a competitive advantage if your skills allow you to code by hand and know and understand theory, while also being able to take advantage of DataRobot and really accelerate your productivity in applying those skills.
I am a customer facing data scientist with DataRobot. Before working here, I used more conventional tools like R and Python to build predictive models. While you can learn a lot about data science working that way, the burden of writing and debugging code can take up a lot of time and headspace, slowing the learning about actual data science principals. I think a lot of people consider writing code a "guardrail" to prevent people who don't know what they are doing from doing data science. The fact of the matter is, this isn't a guardrail - it is a barrier to entry. Even seasoned data scientists who build model in R and Python can make mistakes. Just because you can write syntax and get the code to run, doesn't mean your model is good.
I personally have learned a lot more detail about many different algorithms just by using the software. Imaging how much you will learn when you have 10 algorithms for your specific dataset to investigate? All of the models in DataRobot are transparent and easily tunable, allowing you to try out different strategies and examine the result. You can learn a lot more about the actual modeling, as well as try out many more partitioning/model framing strategies .
One of the things I tell my customers is that you can put a well structured dataset into DataRobot, press start and get very good models. However, the platform is also very deep. You can connect to it using an API with R or Python and do some very sophisticated data science very quickly. We also update our software on a weekly basis, if there is a new algorithm or approach out there, we probably have someone working on getting it into the product.
Finally, I just want to say that it is good you are thinking about these things early in your career. Being concerned about how much you are going to learn is a great attitude to have; regardless on how advanced you are.
I could see if I could get you a free trial of the platform if that will help inform your decision. Would you be interested in that?