A guide to choosing the right data science position
I have been interviewing people from all around the world on my podcast So you want to be a data scientist?. My guests have been from very different positions who all have either the title data scientist or similar. Inspired by everything I’ve heard from my guests on the podcasts, I prepared a list of possible positions you can consider when planning your future.
It’s important to know which one you want to work at, in order to get the most satisfaction out of your job. I understand this is not very easy to do when you have little to no experience in a field. In data science, you can end up in many different types of positions. They all have varying responsibilities, different careers and busyness. But if you are not in the type of position you enjoy, you might end up unhappy.
I didn’t think about this much when I first started. For me, it was a trial and error approach. I have started in a big company as a consultant, only to realise that wasn’t for me. In time I realised an in-house position would suit me better. That’s why I went for a change and I will be starting in my new position next week. I’m happy and excited to experience how being an in-house data scientist will fit me.
If you don’t want to spend a couple of years of your life trying and erring as I did, here is a list of common types of positions for data scientists. You can find the link to the episode where I interview the guest with that position below the titles.
Consultant data scientist
Typical responsibilities: Being a consultant means that you will be working with clients of your employer. Your assignments are projects on client companies. You will be doing data science either in a team or alone in a different location/team than yours.
There are typical responsibilities of a consultant other than the technical data science tasks. These are mostly, helping with coming up with a business scope for the project, communicating with the client about what you’re doing and why it’s important, talking to business-people periodically to update them. It is a very “soft-skills” heavy position.
Projects: Projects are begun by salespeople of the consulting company going around pitching clients. You can get any sorts of projects. By this I mean, they can be in any industry. Think of the energy industry, hospitality, banking, finance, travel. Of course, the industry will depend on which companies your main employer is serving. It sounds interesting to have all the options but it is also likely that for every project you enjoy working on you might get assigned to a couple of projects that you won’t.
The projects might also be at any level. Even if you want to work on NLP, you might be stuck working on making dashboards months at a time. So make sure you understand what type of projects are done in a company before you start working there if you don’t want to be disappointed.
Career advancement: A consultant has both business and technical skills. So you have the option to grow into a managerial role or a more technical role.
Working hours and stress: Working hours can get challenging in consulting. You are not working directly for your employer but a client and consulting companies want to look good to their clients. This means that you might get extra pressure from your employer when it comes to deadlines.
What might have been a soft deadline in an in-house position, might become a hard deadline in a consulting environment just because your company promised the client to deliver at a certain time. This might cause a high-stress environment. But still, some companies do a good job managing this stress. Though I get the impression that working long hours and being fine with stress is “expected” of consultants.
Pros - projects in a variety of industries, mostly good pay, you get to learn the business side of things, actively using your soft-skills
Cons - you might get stuck with the type of projects you don’t like, working hours might be longer than average, the stress of working with a client
In-house data scientist
Typical responsibilities: When you work as an in-house data scientist, you are solely interested in what’s going on inside your company. You will likely be in a team of data scientists and machine learning engineers. The projects will come to you from other teams inside the company. You likely will be in the same office (if that ever happens again that we return to the office) and that will be your main working spot. Apart from working on new projects, you are responsible for maintaining the projects you’ve delivered.
Projects: Projects are started by other teams in the company coming to your team with a request, your team coming up with an idea for a new feature or the need for an update to the currently-in-use systems.
Working hours and stress: Depends on the company. It’s likely that if the company is in a high-pace, high-pressure industry like finance/investment banking, you might work long hours and will be under stress. Though many companies are paying extra attention to not stressing their people out unnecessarily. Make sure you know what you’re signing up for before you start somewhere.
Career advancement: An obvious career path for an in-house data scientist is to become a team leader, manager or even CTO in the given company. The career path lies mostly on the technical side.
Pros - working in the same team, many opportunities to improve yourself especially if your colleagues are experienced
Cons - all your projects will be in the same industry (this might be a pro if you like working in that industry), apart from developing new projects you need to dedicate part of your time to maintaining old projects which might get tedious
Freelance data scientist
Typical responsibilities: The awesome thing about being a freelance data scientist is that your responsibilities are whatever you want them to be. You can choose to work for projects in the industries and the levels that you want. Mainly you would be developing projects with your clients, doing the technical work, presenting the results. Depending on what you agree on with your client, you might do less or more.
Projects: Projects happen by you finding clients and offering your services. You can find clients online, reaching out to your network or follow up on people/companies you know can use some data science support. It is likely that a seasoned freelance data scientist will have a clientele she/he repeatedly works with.
Working hours and stress: You can decide how much and how little you work as a freelance data scientist. This sounds like a dream but of course, it comes with a cost. The less you work, the less money you make. Especially if you’re just starting out, finding projects and clients could be a source of stress.
Career advancement: You can choose to join a company as a data scientist after having a successful career as a freelancer. Or you can decide to become a jewellery designer, or anything else. You have the freedom to choose your own path. Jokes aside though, you can grow your career by becoming an expert in a certain practice/industry or technology. (e.g. deep learning, conversation projects expert, NLP, image recognition etc.) So clients will know to come to you when they need help on these specific domains.
Pros - work whenever, where ever you want, choose the clients/projects you want to work with/on
Cons - the pressure of finding clients, the pressure of delivering to clients on time, you need to be the data scientist, accountant, salesperson, web developer and many other things in your own little enterprise
Researcher data scientist
Typical responsibilities: Out of all the people I’ve had on the podcast, researchers by far are having it the best in my opinion. But of course, it is not easy to get to where they are. Researchers mainly work towards a research goal and they deliver working prototypes and/or papers about the work they’ve done. They work on interesting projects that investigate new ideas and try to push forward today’s technology.
Projects: You can come up with your own research project or join on-going bigger projects. Of course, in order to do your own project, you need funding and that means you need to convince someone of the value of the project.
Working hours and stress: From what I’ve heard, researchers do have some stress over their work. When you’re working on a novel idea, there is a lot of uncertainty after all. You do not know if the project will succeed or if the work you put in will be used in real life. But other than that the working hours and the working environment is pretty relaxed.
Career advancement: If you work at an academic institution as a researcher you have the option to advance academically or join the industry. If you already work in industry or in an independent research lab, you can advance to take on more responsibilities.
Pros - very interesting work, the possibility of working on a passion project
Cons - might not be satisfying for people who want to see their work applied to real-life immediately
Free data scientist
Typical responsibilities: There are not many responsibilities of a free data scientist. This is not a formal position but rather something you can do with your data science skills. A free data scientist would be contributing to open source projects, starting projects by him/herself or with some friends and trying new things and creating new projects out of thin air. It’s very fun to work this way if you don’t have to worry about money urgently. It will get you a lot of experience. And if you want it, make it easier for companies to trust your skills for when you want to be hired in the future.
Projects: Whatever you want to work on.
Working hours and stress: Again, you can work as much or as little as you want.
Career advancement: after contributing to open source projects and creating a couple of yours, companies might already come to you with offers. You can also start working as a freelance data scientist.
Pros - You have unlimited freedom to work on your own projects and spend your time as you will, great experience to show your capability and skills
Cons - Hard to make money this way
Positions that are not titled data scientist but are actively using data science
Typical responsibilities: This is a more general umbrella position. This can be anything from marketing to product development to customer relations. There are sometimes positions where people are not titled data scientist but their main tool is ML and data analysis and they effectively are doing data science. I’ve seen people take positions like that when they realised that their passion lies with the work and not with data science which is just a tool at the end of the day.
Career advancement: If you want to become a data scientist later in life, a position like this can be a great stepping stone in terms of experience but also in terms of understanding if you really want to be a data scientist.