How to Write the Perfect Data Scientist Resume

Need to write a knockout data scientist resume to land your dream job? You’re in the right place. As data scientists, we often obsess over numbers and conversion rates… and that’s a good thing! A job search is just a numbers game with three steps:

Applications ⇒ Interviews ⇒ Job Offers

Today, we’ll look at how you can improve your rate of Applications ⇒ Interviews by writing a winning data scientist resume.

We’ve compiled our favorite tips for writing the perfect data scientist CV, and they’re broken into 3 sections:

• Before Writing Your Data Scientist Resume
• While Writing Your Data Scientist Resume
• Editing Your Data Scientist Resume

BEFORE WRITING YOUR DATA SCIENTIST RESUME

Resumes are often misused as a “credential dump,” a hodge-podge of skills and experiences. Instead, your data scientist resume should tell a persuasive story with YOU as the protagonist.

Each section should work in harmony and each bullet point should add colorful details. That’s why the prep work before writing your resume is even more important than what you type.

In this section, we’ll craft a thoughtful, effective strategy that will help you tell that persuasive story.

1. Step Into the Employer’s Shoes

Chess grandmasters step into their opponents’ shoes, trying to anticipate every move… Expert salespeople step into their prospects’ shoes, trying to understand their pains and desires… And we use this same tactic to write the perfect data scientist resume.

Here’s the thought experiment:

If you were the employer, what would you be looking for in a candidate?

Now, if the answer to that question sounds easy, you’re probably not thinking carefully enough. For example, “smart programmer with data and machine learning skills” is not detailed enough.

Instead, pretend you need to write a complete job description… How exactly would a data scientist help your business? Where would they fit in? Who would they work with/under? Which skills would you require? Which experiences would impress you?

Answering these questions will help you shape a coherent strategy right from the start.

To-Do List:

1. Carefully read the job descriptions for 5 positions you like.
2. Note the patterns. Which skills and experiences show up repeatedly?
3. Note the people. Will you work with the engineering team or with the analysts?
4. Write one or two paragraphs summarizing the perfect candidate in your own words.
5. Re-read this description each time you sit down to write or edit your resume! This will put you in the right frame of mind.

2. Craft an Elevator Pitch

An “elevator pitch” is a short sales pitch that summarizes a product, service, or business. The term originates from the scenario of pitching something in a short elevator ride.

But wait, you’re not on an elevator… right?

Well, even though your resume can be longer, this exercise works wonders when paired with the previous one. It forces you to prioritize the most important credentials to communicate.

Craft a brief pitch for yourself in plain English. In this pitch, succinctly describe your background, highlight key credentials, and then emphasize what makes you unique.

You will then use your resume to convey this pitch (and back it up with details)!

To-Do List:

1. Take inventory of all your relevant experiences, skills, and credentials. We suggest a simple brain dump onto scratch paper.
2. Re-read your summary of the perfect candidate from the previous tip.
3. With that description in mind, consider: Which of your skills and experiences should you highlight? Which ones make you proud? Which ones make you unique?
4. Write a 3 to 5 sentence elevator pitch. “Sell yourself” by prioritizing the most important credentials.
5. Re-read this pitch each time you sit down to write or edit your resume!
6. (Bonus) If you did the brain dump from step 1, you should save this as a reference when writing your resume.

3. Keep Your Eyes on the Prize

Your resume is meant to land you an interview, not secure the job. This mindset shift will help your prioritize.

Be extremely selective in what you include. Start with your highest-impact skills and experiences.

Think about a great movie trailer – it inspires you to watch the movie by showcasing the most exciting or dramatic scenes, skipping everything else.

Now, let’s write a “movie trailer” to convey your elevator pitch!

To-Do List:

1. Based on your elevator pitch, list the 3 to 5 key skills you’ll need to highlight in your resume.
2. Jot down some notes and preliminary bullet points about each one.

For example, let’s say a key skill is “Applied Machine Learning.” You can jot down some notes such as:

• Trained a random forest to predict wine quality.
• Engineered demographics-based features to improve sales models.

No need to finalize every detail or refine the language just yet. Instead, just brainstorm specifics to support your elevator pitch.

4. Find a Simple, Clean Template

A simple, clean template is almost always better than a fancy one.

Your template should “get out of the way” of your story (and be easily scanned by automated screeners like ATS).

Here are free resume templates provided by Hloom. We recommend those in the ATS-Optimized, Clean, or Modern collections.

WHILE WRITING YOUR DATA SCIENTIST RESUME

You’ve planned your strategy, crafted your story, and picked a clean template. Let’s now dive into tips for writing your resume.

First, fill your template with your “movie trailer.” Write down your most impressive experiences and then list preliminary bullet points underneath. Then, fill in all the other sections such as your contact info, education background, skills, etc.

Now you have a quick “version 0.1” of your resume. It’s not ready to be submitted, but at least it has all the pieces. The rest of these tips will help you get it into fighting shape.

5. Build a “Resume Master”

Should you customize your resume for every employer?

Well, first of all, customizing your resume is important because not all of your experiences will be relevant for every position you apply for. And even for one experience, you could write different bullets that highlight different skills depending on what the employer’s looking for.

With that said, you’ll need to balance quantity and quality. After all, this is a numbers game, and you may sometimes need to act quickly before an opening gets filled.

That’s why we suggest building a “resume master” that is about 2 pages long (or twice your final length). In it, you’ll include ALL of your relevant experiences and a variety of bullet points for each one.

Then, for each application, you can quickly “spin off” a custom version by curating experiences and deleting bullet points.

To-Do List

1. Write down ALL of the relevant work experiences / courses / awards / projects that might make it into a submitted resume.
2. Under each experience, write and refine TWICE the number bullet points that you’ll need. For example, let’s say you (a) built a data pipeline, (b) optimized prediction models, (c) analyzed user data, and (d) made presentation decks during an internship. Even if you’ll only keep 2 bullet points for that experience, write bullet points for all 4 tasks for now.
3. For each application, spin off a custom version by keeping the most relevant bullet points. Also, except for special circumstances, try to limit your submitted resume to 1-page max (i.e. be selective and curate!).
4. (Bonus) To save even more time, you can re-use spin-offs for similar jobs!

6. Don’t Bury the Lede

In journalism, the “lede” is an intro that entices readers to continue. “Burying the lede” is not starting with the most important, interesting, or attention-grabbing elements of a story.

Resume reviewers will be scanning, and they might be tired. Your resume could be their 50th of the day.

Do them (and you) a huge favor… Make their job easier!

To-Do List

1. By Section: Structure your resume to put the most impressive sections first. For example, if you’re still in school and have cool course projects (but less work experience), put the coursework section before the work experience section.
2. By Experience: Within in section, you’ll usually list experiences chronologically, but there will be some tiebreakers. For example, you should start with your most impressive course projects in the coursework section.
3. By Bullet Point: Under each experience, the first bullet point should be the most impactful. It should entice the reviewer to stop scanning and read.

7. Write Concise Bullet Points

“Brevity is the soul of wit.”

~ William Shakespeare

Be ruthlessly concise when writing your bullet points. Allow your resume to reflect your communication skills.

• BAD: Trained and optimized a random forest machine learning model on a dataset to predict the quality of wine with an accuracy of 99%.
• GOOD: Trained a random forest to predict wine quality with 99% accuracy.

First, in this context, “trained” (i.e. building the model) already implies “optimized” (i.e. maximizing performance). It also already implies using a dataset so you should remove that detail if you’re not using it to highlight the size of the dataset.

Second, a random forest is already a type of machine learning model. You wouldn’t say “I just bought an iPhone Apple smartphone.”

Finally, verbose phrases like “the quality of wine” or “an accuracy of 99%” can be rearranged into “wine quality” and “99% accuracy” respectively.

To-Do List

1. Read these simple principles for writing clear, concise sentences.
2. Remove implied or redundant phrases.
3. Rearrange verbose phrases.
4. Replace inflated words. For example, use “help” instead of “facilitate” or “manage” instead of “administrate.”
5. Don’t stress about this. Your resume doesn’t need to be a work of art… it just needs to be clear.

8. Include Concrete Metrics

Show, don’t tell.

Numbers and metrics allow you to show the impact you’re able to deliver, instead of relying on vague superlatives.

• BAD: Significantly improved conversion rates and drastically reduced bounce rate using an SVM.
• GOOD: Doubled conversion rates from 0.5% to 1% and reduced bounce rate by 18% using an SVM.

Specific details also make you more memorable! Remember, your bullet points should illuminate your story.

To-Do List

1. Wherever possible, replace superlatives with concrete metrics.

9. Highlight Past Projects

Again… show, don’t tell.

Projects add color to your resume. You can include projects from work, capstone projects from courses, or even side projects that you do on weekends.

This is especially useful for candidates with less work experience. Projects serve as proof of competence in a way that vague superlatives do not.

• BAD: Gained extensive experience with neural networks for image processing, computer vision, and object recognition.
• GOOD: Trained a neural network on 5000 images to classify hand-written digits with less than 0.01% error.

Just as with concrete metrics, specific projects make you more memorable (especially fun and interesting ones). They’re also easy talking points during the actual interview.

To-Do List

1. Jot down all of the relevant data science and machine learning projects you’ve completed.
2. For each bullet point, consider whether describing specific project would better illustrate your strength.
3. (Bonus) If needed, you can even create a new section called “Independent Projects” and list your projects similarly to work experience.
4. (Bonus) Here are some fun machine learning project ideas.

10. Contextualize Academic Coursework

If you decide to include academic coursework, don’t just list courses and grades. Curricula and naming conventions vary from school to school.

Many rejected candidates blame their lack of experience, but sometimes they’ve just failed to communicate the experience they already have.

• BAD: Got an A in CS159: Natural Language Processing.
• GOOD: Wrote Selenium web crawler to scrape and parse unstructured text data for NLP algorithms.

Again, it’s worth repeating the tip from above: Have a resume master version that you can cut down for each submission. You may have learned many things in CS159, but it’s better to describe the 1-3 most relevant projects than to list them all.

To-Do List

1. If you currently simply list the courses you’ve taken, restructure that section to have space for details (it’s fine if you only include the 2-3 most relevant courses instead of all of them).
2. For each course, add context by describing a project or at least listing the specific topics that were covered.

EDITING YOUR DATA SCIENTIST RESUME

We’re onto the home stretch! If you’ve been following along, you should have a “resume master” with relevant experiences, strong bullet points concisely written, and plenty of projects, concrete metrics, and colorful details.

Now, we’ll refine it one more time, which will push your resume over the top.

By the end, you’ll be ready to “spin off” excellent resumes tailored to each position (see Tip #5 – Build a Resume Master).

Let’s jump in!

11. Use Action Verbs, Don’t “Utilize” Them

It’s common advice to start your bullet points with an action verb such as the ones in this list.

However, it can be easy to get carried away… Beware of long and awkward sounding verbs.

Here’s the rule of thumb: If two verbs have roughly the same meaning, choose the simpler one (e.g. “use” instead of “utilize”).

• BAD: Utilize PCA to reduce dimensionality of financial datasets.
• GOOD: Use PCA to reduce dimensionality of financial datasets.
• GOOD: Reduce dimensionality of financial datasets using PCA.

By the way, for those curious, the difference between “use” and “utilize” is subtle.

…for some writers utilize still connotes something more than use, i.e. the implication that a resource has been turned to good account, and used in a profitable, effective or ingenious way:

They utilized water from a nearby stream to cool the engine.

This subtle extra dimension of utilize is unfortunately jeopardized by pretentious use of it elsewhere.

~ The Cambridge Guide to English Usage

Remember, resume reviewers will be scanning, so don’t trip them up with unnecessarily long verbs regardless of subtlety.

To-Do List

1. When appropriate, start bullet points with a clear, purposeful action verb.
2. Read each bullet point aloud to yourself. Mark the ones that sound awkward or pretentious.
3. Replace marked verbs with simpler ones. Common examples include:
   • “improve” instead of “ameliorate”
   • “lead” instead of “orchestrate” or “spearhead”
   • “start” instead of “commence”
   • “help” instead of “facilitate”
   • “save” instead of “conserve”
4. If needed, powerthesaurus.org can help you find simpler synonyms.

12. Translate Technical Jargon

It can be tempting to vomit technical jargon onto data scientist resumes. Of course, the profession is highly technical, so you will need to include certain keywords, algorithms, etc.

However, most HR reviewers are not as technical. While they do look for key skills, you can make their job much easier by using plain English wherever possible.

So, what should be translated? Well, use the job description as a benchmark!

If the job description is highly technical, i.e. filled with jargon, you’re fine staying more technical. On the other hand, if it only lists broad requirements such as “scripting languages” or “machine learning packages,” then cut down the jargon.

• BAD: Reduced prediction error by 5% by trying LASSO, Ridge, and Elastic-Net Regressions and switching to Area Under ROC Curve.
• GOOD: Reduced prediction error by 5% using regularized regression and adopting an error metric more robust to class imbalance (AUROC).

Remember, your resume should tell a story and be the “movie trailer” to land you an interview. Most of the time, you can skip the nitty-gritty details.

To-Do List

1. Translate any bullet points that are drowning in keywords and jargon.
   • Rule of Thumb #1: if you have a standalone, capitalized term, e.g. Area Under ROC Curve, you may need to translate it, e.g. error metric more robust to class imbalance (AUROC). You can still include the keyword in parentheses if it helps clarify.
   • Rule of Thumb #2: if you list several related keywords, e.g. LASSO, Ridge, and Elastic-Net, they can be simplified, e.g. regularized regression.

13. Replace “Scarecrow” Adjectives

Adjectives like “strong” or “experienced” are scarecrow adjectives… they sound good at first, but they offer no real substance upon second glance.

Instead, these types of adjectives lull you into laziness, and they are the antithesis of Tip #8 – Include Concrete Metrics.

• BAD: Strong knowledge of SQL and experience writing complex queries on raw data.
• GOOD: Wrangled AWS log data and built pipeline to front-end dashboard using SQL.

Again… show, don’t tell.

To-Do List

1. Scan your resume for scarecrow adjectives like strong, experienced, capable, or deep.
2. Replace the scarecrow with concrete metrics or rewrite the bullet point as a specific example.

14. Curate, Curate, Curate

When it’s time to “spin off” a resume for submission, think like a curator compiling a beautiful collection from all the pieces in your “resume master” (See Tip #5 – Build a Resume Master).

For example, don’t try to showcase all your skills. Instead, prioritize the ones most relevant to the position.

Think about restaurants – those with fewer menu items (such as high-end restaurants with prix-fixe menus) are usually perceived to be higher quality!

To-Do List

1. Before you spin off a resume from your resume master, re-read the job description. Curate your bullet points to showcase skills that are most relevant for the position.

15. Give Your Resume to a Friend

When learning data science, a valuable test is to “explain an algorithm to a friend with no background in machine learning.” This forces you to know the intuition behind your tools.

This final exercise is similar. Give your resume to friends who aren’t data scientists. They should still able to identify your key credentials, even if they don’t understand all of the terms.

To-Do List

1. Give your resume to friends who aren’t data scientists. As them what they think are your key credentials. See if they understand your story.

Summary

Your resume’s only one part of your application… but without a winning resume, the rest won’t matter. To recap:

Before Writing Your Data Scientist Resume

1. Step Into the Employer’s Shoes
2. Craft an Elevator Pitch
3. Keep Your Eyes on the Prize
4. Find a Simple, Clean Template

While Writing Your Data Scientist Resume

5. Build a Resume Master
6. Don’t Bury the Lede
7. Write Concise Bullet Points
8. Include Concrete Metrics
9. Highlight Past Projects
10. Contextualize Academic Coursework

Editing Your Data Scientist Resume

11. Use Action Verbs, Don’t “Utilize” Them
12. Translate Technical Jargon
13. Replace “Scarecrow Adjectives”
14. Curate, Curate, Curate
15. Give Your Resume to a Friend