The Difference Between Junior, Mid-Level, And Senior Data Scientist Jobs
When looking for a data science job, you will have to chose what job seniority to apply to. There are junior data science jobs, there are mid-level data science jobs, and there are senior data science jobs. An email subscriber recently asked for how they should think about the different levels.
How to think about data science job levels
One way to think about it is the following: how long you could leave someone alone to do a task without checking in. Rather than thinking about specific skills, different tool sets, years of experience, level of education, subject expertise, maturity level, or anything else, the best way to think about it is holistically. Thinking about it by how long you could leave someone alone to do a task without checking in will help you figure out what job seniority you should be applying for.
Example project and example thinking about different roles
Let's say you have been tasked with building a predictive model that your company will rely on. It needs to be delivered in 6 months. It involves combining external and internal data sets, it involves putting the data sets together, it involves fact finding for requirements, it involves working with various other groups in the company, it involves working with the company's biggest customer, it involves mathematical & statistical modeling and it has to work.
Which comes back to how long would you let someone tackle this project without checking in.
Using the question (“how long you could leave someone alone to do a task without checking in?”), we can the describe the senior, mid-level, and junior data scientist roles as follows:
- Senior Data Scientist - You give it to them and they take care of it. No need to check in.
- Mid-level Data Scientist - You check in probably weekly or monthly but don't get involved too much. You also pair them with a Senior Data Scientist.
- Junior Data Scientist - You check in daily if not twice daily and you pair them with a mid-level and senior data scientist.
Which means, using the time-based classification above
That maturity, specific skills, tool sets (industry and company specific), years of experience, education level, and everything else all matter equally. Because each one will have to come into play with all the various moving parts of the 6 month project. So you can use this to think about yourself or roles you are applying to in order to figure out where you fit and at what level you think the job description is really looking to hire.
It also means that though the definitions can be fluid from company to company, if you think about it in the above terms, it's relatively easy to place yourself or others. After all, if the money is on the line, how comfortable / confident would you be leaving someone by themselves.
What about PhDs and PostDocs?
The email subscriber then asked us why it appears that PhDs and PostDocs can get to senior level faster. As a side question, this also helps to answer why PhD degrees and PostDoc positions are generally wanted by people hiring Data Scientists. This is because doing a PhD, forces you to do all of the above mentioned work by yourself, so maturity is developed.
Here are three main things that are true of PhDs and PostDocs...
- PhD degrees are generally 4-6 years of hard research experience where the person has to be internally motivated and has to deliver results. No results delivered, no PhD degree given. So although they aren't a Data Scientist, they have 4-6 years experience driving projects by themselves.
- Doing a PhD (for almost 99%) of students, involves amazing amounts of failure and then being forced to figure it out. So a PhD student will be resourceful.
- Doing a PhD requires you to do something novel (no matter how insignificant it might seem to non-academics), the whole point of doing a PhD is novel research and novel results.
Ability to Get Sh*t Done is important
Last, but not least, the difference between junior, mid-level, and senior data scientist often comes down to how much knowledge / experience they have getting sh*t done. A junior person may come up to walls and not know how to get around them (either institutional, tools, methods, data, math, statistics, etc). A mid-level person will be somewhat battle hardened and have a pretty good feel for how to get things done and what "short-cuts" to take and how to shoe-horn one method into another one. A senior data scientist gets things done. Period. That is what they do - they will get the project done come hell or high water. Whether that means hiring new people, bringing in consultants, contractors, researchers, academics - they know what needs to happen, they know how to make it happen, and they do in fact make it happen.
So when thinking of data science jobs, ask yourself the question
As you decide what jobs to apply to and what job seniority you think you fit into, it'll be helpful to ask yourself: "how long could someone leave me alone to do a task without checking in". Be honest with yourself. Honesty will help you find the perfect role and the perfect level. Yes, of course strive higher and have stretch goals, but keep in mind that those interviewing you will be doing the calculation themselves as well when evaluating you as a candidate.
To that end, start developing the start to end skills and thinking required for projects. Think about what you would need, what obstacles you would face, and what you would need to overcome them. Then as you work on your data science portfolio projects, you get a feel for the rhythm of doing projects!
Good luck!