Alright y'all, let's talk about how to write a data science resume coming from academia
I'm not going to pretend to be the final word on what a good resume is, and I've noticed some themes while trying to help people
Let's ride
I'm not going to pretend to be the final word on what a good resume is, and I've noticed some themes while trying to help people
Let's ride
My high-level advice is:
1. Get the resume onto 1 page
2. Give yourself credit for progression
3. Show don't tell re: coding + analysis
More details for how to do each of those + resources in the rest of the thread!
1. Get the resume onto 1 page
2. Give yourself credit for progression
3. Show don't tell re: coding + analysis
More details for how to do each of those + resources in the rest of the thread!
1. Get the resume onto 1 page
Folks coming from academia have a lot of fantastic experiences, and the success of a CV is primarily measured in length
Not here
Some application systems will cut off anything after the 1st page, and recruiters spend ~5-20 seconds per resume
Folks coming from academia have a lot of fantastic experiences, and the success of a CV is primarily measured in length
Not here
Some application systems will cut off anything after the 1st page, and recruiters spend ~5-20 seconds per resume
What should you drop to get onto 1 page?
- Summary/mission statements (your experiences should speak to this)
- Relevant courses (folks don't care what you took, they care what you can do)
- Positions not related to DS (you can wow them with clinical experience in the interview)
- Summary/mission statements (your experiences should speak to this)
- Relevant courses (folks don't care what you took, they care what you can do)
- Positions not related to DS (you can wow them with clinical experience in the interview)
If you're having trouble getting your resume onto 1 page, I'd recommend using this template by @GergelyOrosz
It was originally designed for software engineers applying to tech jobs, and its simplicity + focus on deliverables ports well to Data Science
blog.pragmaticengineer.com/the-pragmatic-…
It was originally designed for software engineers applying to tech jobs, and its simplicity + focus on deliverables ports well to Data Science
blog.pragmaticengineer.com/the-pragmatic-…
2. Give yourself credit for progression
Resumes from academia can make people look like they were "stuck" at one level for a long time when nothing could be further from the truth
The lack of formal title progression inside academic structures can hold folks back
Resumes from academia can make people look like they were "stuck" at one level for a long time when nothing could be further from the truth
The lack of formal title progression inside academic structures can hold folks back
Translate your progress into something a data science recruiter can understand
Generally, I recommend PhD students "level" their first 2 years as Data Analyst, the next 2 years as Senior Data Analyst, and the final year(s) of their PhD as Data Scientist
Generally, I recommend PhD students "level" their first 2 years as Data Analyst, the next 2 years as Senior Data Analyst, and the final year(s) of their PhD as Data Scientist
These titles accurately reflect the progression of what a lot of PhD students looking to move to data science positions have experienced!
Of course, this is only one example, and please don't do anything that you feel doesn't accurately reflect your progression
Of course, this is only one example, and please don't do anything that you feel doesn't accurately reflect your progression
But don't sell yourself short!
Data science recruiters operate with limited information + time. If your most recent title is "Clinical Psychology PhD Candidate" they won't be able to make the translation on their own
Data science recruiters operate with limited information + time. If your most recent title is "Clinical Psychology PhD Candidate" they won't be able to make the translation on their own
People perform an absurd variety of roles in academia, give yourself the most relevant title for the job you're applying to
3. Show don't tell re: coding + analysis
Under each of your levels, you'll want bullet points (This isn't the only way to do the resume, but I've come around to it!)
Each bullet will focus on the data science-specific value you've created using skills employers want you to have
Under each of your levels, you'll want bullet points (This isn't the only way to do the resume, but I've come around to it!)
Each bullet will focus on the data science-specific value you've created using skills employers want you to have
A lot of folks are wise to the idea of using action verbs + quantifying things
The problem tends to be they talk about technical quantities ("Ran ML algorithm with an R2 = 0.48") instead of business-value quantities ("Reduced churn from a digital mental health product by 8%")
The problem tends to be they talk about technical quantities ("Ran ML algorithm with an R2 = 0.48") instead of business-value quantities ("Reduced churn from a digital mental health product by 8%")
I recommend people start each bullet point with the real-world value the analysis or project created
Then show which tools + techniques you used to create that value
Then show which tools + techniques you used to create that value
A concrete example:
Instead of "Ran multilevel models on over 80,000 participants to quantify changes in psychopathology and published 7 papers"
Try "Delivered 7 mission-critical reports to technical stakeholders based on multilevel analyses conducted using brms in R"
Instead of "Ran multilevel models on over 80,000 participants to quantify changes in psychopathology and published 7 papers"
Try "Delivered 7 mission-critical reports to technical stakeholders based on multilevel analyses conducted using brms in R"
How to exactly quantify + translate value will differ based on field, so brainstorming with #AltAcChats who share your background can help!
You can also follow @AltAcChats for more resources. Their conversations extend beyond data science, so you can also explore other paths
You can also follow @AltAcChats for more resources. Their conversations extend beyond data science, so you can also explore other paths
Conclusions/Reflections
I hope this is a helpful resource for folks! I'm sure some folks will have different opinions, and that's great + expected
Resume evaluation is often a bit of a black box, and I don't want to pretend to have any silver bullets
I hope this is a helpful resource for folks! I'm sure some folks will have different opinions, and that's great + expected
Resume evaluation is often a bit of a black box, and I don't want to pretend to have any silver bullets
Also, my resume looked like this when I got my first data science job
An updated version is in the works, and you'll notice this version doesn't meet a lot of the standards I set out in this thread!
mcmullarkey.github.io/resume_markdow…
An updated version is in the works, and you'll notice this version doesn't meet a lot of the standards I set out in this thread!
mcmullarkey.github.io/resume_markdow…
How do I reconcile that?
I'm trying really hard not to give advice solely from a "just do what I did" perspective
That resume got *me* a data science job, and I had a lot of non-generalizable advantages
I'm trying really hard not to give advice solely from a "just do what I did" perspective
That resume got *me* a data science job, and I had a lot of non-generalizable advantages
I know "just do what I did" still infects my advice
And several things make me pretty sure my exact path to a data science job isn't replicable (My positionality + privilege, competing well in a live-streamed ML competition, being mildly Twitter-known, probably other things!)
And several things make me pretty sure my exact path to a data science job isn't replicable (My positionality + privilege, competing well in a live-streamed ML competition, being mildly Twitter-known, probably other things!)
Still, that old resume is proof you can get a data science job without following this advice!
I'm just doing my best to smooth the way for folks who don't have all the (unearned) built-in advantages I do
Open to good-faith feedback, have no time for "why leave academia??"
I'm just doing my best to smooth the way for folks who don't have all the (unearned) built-in advantages I do
Open to good-faith feedback, have no time for "why leave academia??"
• • •
Missing some Tweet in this thread? You can try to
force a refresh
