DECODINGDATASCIENCE
JOBDESCRIPTIONS
Co-organizer of PyLadies Hamburg


Board member of Python Software Verband


D&I and CoC of Python Software Foundation


Art: tiyepyep


Past: data science lead/manager, data
scientist, data engineer ..
TEREZAIOFCIU
HEAD COACH DATA SCIENCE @NEUEFISCHE
HOWITSTARTS
Lots of people want to work as data
scientist


Many jobs are open and companies are
looking for all sorts of data scientists
TEREZA IOFCIU
HOWITENDS
Most of the times there is a huge
mismatch between expectations
and reality


Both for companies and employees
TEREZA IOFCIU
GOAL
MISSMATCH
PEOPLE LOOKING FOR JOBS


VS


COMPANIES HIRING
JOBSEEKERSWANT
Understand what role they will have,


where (culture) and with whom (team) will they spend
a lot of their time with
MANYCOMPANIESWANT
Find out with minimal e
ff
ort if a
candidate
f
its the job


Are you smart?


Are you compatible to the culture?
FINDINGTHE
DATASCIENCE
DREAMJOB
EVALUATINGJOBS
Job Descriptions
Interviews
On the Job
Takes least time investment to assess
JOBDESCRIPTIONSTAGE
Doing research at this stage will only help in the future!


Research: job description, company site, linkedin
employees, blogs, meetups


Think of missing information: WHY is it missing?


Ignorance vs Indi
ff
erence: perfect questions to ask in
interview
WHAT CAN WE FIND OUT?
JOBDESCRIPTION
ROLE CULTURE
TEAM
YOUWILLSPEND~30%OF
YOURTIMEWITHYOURTEAM
TEAM
TEAM
Team not mentioned..


Team mentioned as in progress


Team well described
TEAM
Team pro
f
iles online


How diverse are they


What di
ff
erent roles do they have


Do they inspire you (blogs, open source,
meetups, conferences..)
TEAM
Are there any juniors?


How is mentoring happening?
INANIDEALWORLD


COMPANIESHIREPEOPLETHEY
NEEDNOTPEOPLETHEY“WANT”
ROLE
ROLE
Some bullet points


Many bullet poitns


Too many bullet points ->
maturity of the team
24%OFSURVEYRESPONDENTSSAIDTHATTHEY
THOUGHTTHEIRORGANIZATIONWASDATA-
DRIVEN
meanwhile in 85 Fortune 1000 companies
BIG DATA AND AI EXECUTIVE SURVEY 2021
BEFOREDATA
Data is the new oil


You are our
f
irst data scientist
Data science is magic


Immediate results


All our problems solved
Decisions made based on gut
feeling


No support from upper mgmt for
going data driven


Reporting and data
f
lows are still
at a manual level
TEREZA IOFCIU
BEFOREDATA
Assisting with reporting or doing all the reporting


Convincing that data needs to be tracked, collected and
analysed


Most e
ff
ort will be spent on politics


Company needs are: business intelligence and data engineering
TEREZA IOFCIU
LIKINGDATA
We want to be data driven


Data science is still magic


Immediate results and on
demand improvements
Lack of company wide data
culture


Some decisions based on data
insights


Solid data infrastructure
TEREZA IOFCIU
LIKINGDATA
Convincing people that decisions should be backed by
data


Do analysis and modelling, though many models will not
make it live


Lots of e
ff
ort spent on politics and educating others


Lessons learned in prioritising of work


Company needs: data literacy training
TEREZA IOFCIU
25%
DECISION
MAKERS
AREDATA
LITERATE
DATADRIVEN
We publish research


Data is at the core of the product


Complex problems should be
solved by data science


Data is in every product/team
Data literacy in over 50% of the
company


Decisions based on data insights


Open source, research,
publications
TEREZA IOFCIU
DATADRIVEN
Building data products with the team


Advancing the state of the art of research


Advocating for data science / your team / product


You will be doing data science and more
TEREZA IOFCIU
LACKOFCULTUREFITLEADSTO
STRESSANDCANTAKEALOTOF
YOURWORKTIMETODEALWITH
CULTURE
CULTURE
Competitive vs collaborative


Volunteering


Flexibility
CULTURE
Inclusive language


Gender coding


Ableism


Ageism


Listings with gender neutral
wording get 42% more responses
CULTURE
The lists of skills and
requirements... are all marked
as important?


Too many bullet points -> rigid
hiring process and women will
be less likely to apply
CULTURE
Performance objectives


Even if they are missing you can turn the skill list into performance objectives
and explain why you are quali
f
ied


Where would you apply?


a PhD plus 3 years experience
Improve the performance of 3 of
our production ML models in the
f
irst 6 months
CULTURE
Equal Employer Opportunity
Statement


Does it exists?


Is it generic?


Jobs with EEO
f
ill 6% faster
Expect to see it
74 %
Compensation
61 %
PEOPLE WANT TO SEE IT.. COMPANIES HIDE IT
THESALARY
Show it
1 %
Don't
99 %
JOBDESCRIPTION
Ignorance vs Indi
ff
erence


You need to decide which one are you willing to deal with


What is important for you?
WHYDOWE
NEEDTODO
THIS?
DECODING JOB DESCRIPTIONS IS JUST
DEALING WITH SYMPTOMS OF A
BROKEN SYSTEM
MARY BAJOREK (LIFEWORTHLOVING COACHING)
WEARESTILLINTERVIEWINGFORFILLING
INCHAIRSRATHERTHANINTERVIEWING
PEOPLETOPARTICIPATEINATEAMAND
ACHIEVECOMPANYGOALS
TEREZAIOFCIU


@TEREZAIF
https://hbr.org/2021/02/why-is-it-so-hard-to-become-a-data-driven-company


https://www.themuse.com/advice/not-a-culture-
f
it-at-current-job


https://textio.com/blog/how-to-craft-a-sincere-equal-opportunity-employer-
statement/28880187459


https://talentfoot.com/2020/08/04/gender-neutral-job-descriptions/


https://hbr.org/2014/08/why-women-dont-apply-for-jobs-unless-theyre-100-
quali
f
ied


https://www.linkedin.com/business/talent/blog/talent-acquisition/how-women-
f
ind-jobs-gender-report


https://www.linkedin.com/business/talent/blog/talent-strategy/highly-e
ff
ective-
ways-to-eliminate-hiring-bias


https://builtin.com/job-descriptions/how-to-write-a-job-description


https://www.lifeworthlovingcoaching.com
REFERENCES

Decoding Data Science Job Descriptions