Data scientist job description template

A data scientist

JD that is not generic.

One short, opinionated template. 280 words, not 1,800. Below: a side-by-side teardown of the typical job description and what to cut, the template ready to copy, and a one-click button to post it via Picked.

One-click post

Do not want to copy-paste? Skip the document and let us post a role with this template plus your specifics. ~30 seconds.

Use this template →

$0.99 per AI-vetted candidate. First 50 free.

The teardown

Side by side. Cut everything in red.

A typical JD · 1,847 words

scroll-killing

Data scientist (typical posting)

About us · cut

Founded in 2019, ACME Inc is a fast-growing, mission-driven, customer-obsessed scale-up empowering the modern hiring loop. We have been featured in TechCrunch, won several awards, and our culture has been recognised by Best Places to Work.

About the role · cut

You will be joining a passionate, collaborative team of self-starters in a dynamic, fast-paced environment. We are looking for a high-performing professional who is excited by ambiguity and ready to wear many hats.

Responsibilities · cut

- Collaborate with cross-functional partners to align stakeholders - Drive impact across the business - Champion excellence - Be a culture ambassador for the team - Wear many hats - Lots more bullets like this

Requirements · cut

- 5+ years of professional experience - Strong communication skills - Self-starter - Team player - Comfortable with ambiguity - Bachelor's degree or equivalent

Nice to haves · cut

- A laundry list of every tool anyone on the team has ever used - MBA / advanced degree - Previous experience at a high-growth startup

Benefits · cut most of this

- Competitive salary - Equity - Health, dental, vision - Unlimited PTO (which no-one takes) - Free snacks (in office) - Annual offsite - Pet-friendly office - Pinball machine

Picked template · 226 words

ships

Senior data scientist · London or remote

We are hiring our second data scientist. You will own a model from problem framing to production monitoring. The first project is the ranking model that drives our recommendations. The team is one scientist, two engineers, one PM. We ship behind feature flags and read the eval before the demo.

One paragraph, not six sections

Context. What the role owns. Team size. How you ship. Everything else is noise until the interview.

Skip the company overview

They found your JD. They can find your website. They will skim it. They will learn about you in the interview if they are still interested.

Bullets only where they earn their place

A list of "responsibilities" tells the candidate nothing they cannot guess.

Post the comp band

You will filter the wrong people anyway. Either pre-application or after offer. Pre is cheaper.

The template · plain text

Copy. Paste. Replace the bracketed bits.

Post with this · 30s

data-scientist.md

226 words · 3 min read

# [Senior/Mid/Junior] data scientist · [city] or remote

We are hiring our [Nth] data scientist. You will own a model from problem framing to production monitoring. The first project is [the model, the surface]. The team is [N] scientists, [N] engineers, [N] PMs. [How we ship].

## You will be good here if

- You have shut a project down after the EDA because the answer was already there.
- You pick the metric before you see the result.
- [The third trait you actually care about].

## What you will not get

A pure notebook job. You will own the deployed model and the on-call that comes with it. We do not separate ML research from production.

## Compensation

[Currency][low] to [high] base. [low] to [high] percent equity. We post the band because we mean it.

## How we hire

30-second post, screen + assessment via Picked, 20-min interview via Picked, one on-site half-day with the team plus a live modelling design discussion. We aim to give a yes or no in 7 days.

## What is the rubric

Modelling judgement, communication of uncertainty, evaluation honesty, cross-team comms, curiosity. Not the count of frameworks on your CV.

That is the JD. Apply via [link] or just send a paragraph about the last project where you decided not to build a model.

[Your name], [your title]

Tone rules

Eight phrases to delete
from every JD you ever write.

"Rockstar / ninja / wizard"

Nobody respectable applies to this. Replace with: nothing.

"Self-starter"

Means: we have no plan and no manager. Replace with: name the actual autonomy.

"Comfortable with ambiguity"

Same thing again. Replace with: tell them the specific ambiguity they will face.

"Fast-paced environment"

Means: chaotic. Replace with: name the cadence, sprints, ship-day, on-call.

"Champion"

A verb that means nothing. Replace with: the actual outcome.