What Is the Career Path for Data Scientists?
Vložit
- čas přidán 21. 11. 2020
- Subscribe to RichardOnData here: / @richardondata
Patreon: / richardondata
Will Data Science Die in 5 Years?: • Will Data Science Die ...
In this video, I talk about the career path for a data scientist across three horizons: a short term of 1-5 years, a medium term of 5-10 years, and a long term of 10+ years.
Links:
"Data Science is Dead" - Alexs Thompson: / data-science-dead-alex...
"Data Science Career Path & Progression" - Julien Kervizic: / data-science-career-pa...
"Progression of a Data Scientist" - Sequoia: / progression-of-a-data-...
"The AI Hierarchy of Needs" - @mrogati: hackernoon.com/the-ai-hierarc...
"Becoming a Level 3.0 Data Scientist" - Jan Zawadzki: towardsdatascience.com/becomi...
"CIOs Must Lead the Data Science Charge": www.cio.com/article/3400637/c...
"How to Become a CTO" - Maryville University: online.maryville.edu/online-m...
"Can Data Scientists Become CEO?" - Jeremy Wyatt: / can-data-scientists-be...
PayPal: richardondata@gmail.com
Patreon: / richardondata
BTC: 3LM5d1vibhp1F7pcxAFX8Ys1DM6XLUoNVL
ETH: 0x3CfC599C4c1040963B644780a0E62d45999bE9D8
LTC: MH8yPjvSmKvpmRRmufofjRB9hnRAFHfx32 - Věda a technologie
Fellow data scientists: what does your career path look like?
I agree with the following definition: assuming domain knowledge, data science follow a path with more programming than any statistician, and more statistics than any programmer.
- Data science doesn't have a straightforward career path. It's a new field, starting in 2010, and is very broad and varies job-to-job and person-to-person. But data science should be a robust field going forward.
Short-term (1-5 years) = entry-level analyst / data scientist
- First role in DS, won't be exposed to whole DS pipeline until 1-2 years in
- See illustration of the AI heirarchy of needs - this illustrates the data science pipeline well.
- Early on, you'll likely focus on analytics and metrics, maybe some cleaning, AB testing, algos
- Later you might do more with AI/deep learning. If not, you may want to change companies in your first 5 years.
- You should also gain understanding of the data engineering and tasks involved earlier in the DS pipeline
- This broad exposure will give you an idea of what you may want to specialise in later
- A good DS study pathway is to learning SQL then R/Python very well. Then to go back later to learn Python/R well enough.
- As you solve problems, you'll compound your skills in the chosen language. Don't worry as much about knowing a tonne of languages straight away.
- At your current job, you should be able to branch into different technologies (javascript, julia etc) if you can find use-cases.
- Looking at new jobs, you may want to consider how it will let you expand your technical skillset.
- In 6-12 months, you'll get a good understanding of the domain you work in. But over 3-5 years, you'll learn an exceptional amount about your industry (healthcare, finance, automotive etc) and this will make you a much better data scientist and let you add strategic input in your job.
- When you start, you'll be a code monkey, solving problems given to you. But over time, you'll deal with unstructured problems or just situations where you need to find the right problem to solve, then figure out how to solve it and solve it.
Medium (5-10 years) = senior data science/data science manager
- By now you should have a good title and salary that reflects your ability and experience.
- This is also a good time to consider pivoting in your role. Maybe product manager, software engineer. Or stick to data science and become a subject matter expert (e.g. R shiny, NLP etc)
- You'll be managing initiatives/projects, managing people etc
- You won't be the head guy but you will be a thought-leader in your area of expertise
Long-term (10+ years) = looking towards Chief Data Scientist, CTO, CIO etc, even CEO
- Key IT areas: network architecture, big data engineering, information security management, security engineering, web software development. You'll dabble with these things
- 10 years is a very long time period to build experience in a particular career, so you'll have a lot of options
I enjoy watching your videos. I'm an undergrad and I'm very interested in data science. You share some great insights and I always learn a lot thanks!
I'm glad to hear that and I wish you well as you prospectively enter the field!
Richard, you are really amazing, what a beautiful and energetic description! As an Iranian researcher working in Czech Republic, I just say thanks for sharing your experience with the world. 😍🤗
Wow, that's really complimentary. Thanks so much! That's exactly the idea - bring forth my knowledge and research in a blunt albeit energetic fashion.
@@RichardOnData It is totally truth. Your expression is really motivational. Keep doing this priceless job :))
I'm studying a bachelor in biochemical engineering. I see that data science will have a great impact here. I agree with you about the importance of domain knowledge. Thanks for the video.
It really is, and I'm glad that you know this even in the process of getting a bachelors, because many don't fully understand this until much later!
Thanks for sharing so many helpful thoughts
Your views and insights are simply awesome. Thanks!
I appreciate that! I try to keep things as straightforward and simple as possible!
Great Richard, as always. Anyway I would like to get deeper about this topic, because a career last 40 years, and I guess it should be similar to the software engineer's one. Technical skills, in tech, dont last long and change so quickly, that it seems the only way to stay competitive is to go into management, CTO and then someone opens its own company. About the last one, how data science knoweldge/experience would be suitable to start a company? And so, which domain would be the best?
In terms of selecting a domain, I think most people find their way into a domain by mistake. However I know several people personally who have cited the domain as the worst aspect of the job and what makes them want to leave almost more than anything else. For example if you care nothing for cars, probably don't want to get into the automotive sector! I have stuck with healthcare and pharma, personally, because they are the most interesting feeling to me - ultimately you just want to go with something that appeals to you personally.
Now as far as how much knowledge and experience is necessary, that is really going to depend. I know someone who started his company as soon as he got out of college. Probably anywhere from 5-7 years would be a good jumping-off block and would provide enough experience to get an idea for what does and doesn't work in business!
Wonderful insightful video , as always 👍🏽
Much appreciated!
As someone studying data science, this helped validate my choice. Thanks for the video!
Hi I wonder what industry are you working in? I'm now a marketing undergrads but trying to shift my career path by taking a data science master's degree. I hope to become a marketing scientist/analytics, but currently there's not too many info of this position/niche online. It would be great if you are familiar with it!
Great video.
Thank you! It was fun to make!
Hey Richard I'm currently an undergrad student majoring in Information Science. I was wondering if I need to go back to school to get a Masters degree to advance my career. If so, what would be a nice masters degree to compliment my Information Science Degree. Thanks!
Are you going into Data Science / Data Analytics? If you're going into Data Science, then yes, it absolutely couldn't hurt, but a recent very large study found that as many as 40% of the people in data science do not have advanced degrees - so it's certainly not a requirement. My recommendation for complementing an Information Science degree, if you are to go that route, is Statistics.
Do data Science will be ending in late 2020s or 2030s???
Funny: "Let's say you're five years in to data science." Smiles. "Hopefully your title kinda reflects that." Later on, "As well as a handful on generous pay raises!"
Question: Do you think the election was fraudulent, from a data perspective? Are you planning on doing any follow up on that?
Question: What do you think are the self-employment prospects for new and established data scientists?
I have not really been able to judge the results from a data perspective, at least yet. I will allow the proceedings to play out in court, as there are certainly some very intense allegations of fraud - but intense allegations do require intense evidence. There are some county by county findings in states like MI and WI that I find a little puzzling, but from eyeballing these things I also don't think that it's enough to overturn those states let alone an election.
As far as the self-employment prospects, they're fairly strong for data science as a whole but significantly more challenging when you're in the "new" stage. Insofar as my youtube channel is concerned you're talking to a self-employed data scientist right here. :)