Creating the (Im)perfect Developer Interview

GayleL. McDowell | Founder/ CEO, CareerCup
gayle in/gaylemcdgayle
Creating the (Im)Perfect
Dev Interview Process
It’sall broken.Here’showto make it a littleless.
Oct 14, 2015

gayle in/gaylemcdgayleGayle Laakmann McDowell 2
Hi! I’m Gayle LaakmannMcDowell
Author Interview Coach Interview Consulting
<dev> </dev>
(CS) (MBA)

Gayle Laakmann McDowell 3gayle in/gaylemcdgayle
Yes! Slidesare online!
Gayle.com
Click“Events”
Find Event

How People Hire
Current strategies
01 Email: gayle@gayle.com
Subject: #interviewtraining

gayle in/gaylemcdgayle 5Gayle Laakmann McDowell
How
We
Hire
Problem-solving questions
Onsite “real-world” tests
Homework projects
Take-Home Tests
Prior experience
References

It’s all broken
There’s no “perfect” interview.
Always false negatives.
Always false positives.

Problem Solving
Intelligence matters (a lot)
Lots of communication
Ability to push through
hard problems
Intimidating
Prep helps
Other ways to be “great”
Can turn off sr. candidates
Lots of bad questions

Onsite Real-World Tests
See real-world skills Identifiescurrent skill
set, notpotential
Lack of communication

Homework
Sees real-world skills.
Cheap & easy
Identifiescurrent skill
set, notpotential
Candidatescan feel
exploited

Prior Experience
It matters
Lots of communication
Some can “talk a good
game”
 Somecan’t…

References
Prior experience is
good
Hand-picked
Differentcompanies
have differentneeds
Not always 100%
honest

gayle in/gaylemcdgayle 12
z
Gayle Laakmann McDowell
Not
To
Mention…
NONE evaluate workethic or
ability to focus
(whichisn’tnecessarilyconstant
anyway)

z
What
To
Do?
Find what’s right for you
Identify theflaws
Make it as un-brokenas possible
Mix n’ match

Core Beliefs
Philosophies around hiring

#1: Interviews don’tneed to mirror the real world

#2: Goodcandidateexperiences matter

#3: You’llnever have a perfect process

#4: You need to THINK aboutyour process

The Process
The Good, The Bad, The Ugly

A typical good process
Resume Selection
Intro Call w/
Recruiter
Email that outlines
process
Code AssessmentPhone Interview
~4 onsite
interviews
Discussion &
Decision
“Sell” Lunch /
Dinner

Mistake#1: Expectations
Mistake: Candidates don’t know what to expect
 Howmany interviews? Will I be coding? What sorts of questions?
 Leads to false negatives & bad reputations
Fix: You should have a standard “what to expect” email.

Mistake#2: Wasted Phone Interviews
Mistake: Phoneinterview are often:
 Tooeasy
 Totallydifferentfromonsite
Fix: Should match onsite interviews in difficulty &style

Mistake#3: Lack of Structure
Mistake: Interviews are often a “free for all”
 Interviewersask whatever theywanttoask
 Leadstorandomnessinprocess& unpreparedcandidates
Fix:
 Assign“focuses”upfront.
 Traininterviews.

Mistake#4: Not Knowing Your Numbers
Mistake: You don’t know your metrics
 Howwillyouknowtheissues?
Fix: Knowand track:
 Passrate ateachstage
 Timetoget througheachstageofprocess
 Average scorebyinterviewer
 Correlationinphonevs.onsitescores
 Typicalonsitevisits

The Interview
What do you ask and why?

Goals
Great employees, not great interviewees
Good in 6 months,not 6 days
Reduce falsenegatives
Keepcandidates happy

What
Do You
Need?
Soft Skills / CultureFit
Prior Experience
Knowledge
Open-EndedDesign
ProblemSolving

Soft Skills / Culture Fit
Yeah, you probably need it
Email: gayle@gayle.com

z
What
Is
It?
Communicationskills
Teamwork/Leadership
CultureFit

Where it goes wrong…
Not gettingwhat you thinkyou’re getting
 “Areyoumore detailorientedor bigpicture?”
Cheesy questions
 Get cannedresponse
Your culture: notwhat you thinkit is

Ask Yourself
Whatisyour culture exactly?
Can you identifythose things?How?
Do you need to explicitlyidentify these?
… Or is it just “no jerks”?

Best Practices
Define your culture /needs
Identify whatyou need to assess
 (orifyouneedtoassessit)
Avoid cheesy questions
Use experience, not hypotheticals

Prior Experience
What have they done before?

Why?
Isthisapersonyouwanttoworkwith?
Technical expertise
 Hasmadegood,interestingtechnicaldecisions
Culturefit/personality
 Notarrogant,curious,initiative,etc.
Communication
 Can theyarticulateimpact?

Interviewers take answers as-is
 Candidatesunderplayaccomplishments
Overly specific questions

Best Practices
Probe deeper
Be nice and friendly
 (Evenifyoufeeldifferently)
Stick to more technicaldiscussions
ChallengeYOUR assumptions
Inevery interview

Evaluation
A factor in ALL interviews
Err towards listing minor concerns
 Evenifit’sjusta “feeling”
Challengeyour assumptions

Best Practices
Overly specific questions
“Cheesy” canned questions
Accepting answer as-is
 Probedeeper
Many people give TOO MUCH credit to team

Examples
Dive into a technicalproject
 Walk throughdesignonwhiteboard
 Discusstradeoffs,keydecisions,etc
 Extensionstoproject(scaling,etc)
 Focusonpersonalimpact

Technical Knowledge
But do they know stuff?

TechnicalKnowledge
Can be a reflectionof depth
Hires people who can makeimmediate impact

Too surface level
Not a “reflection”of anything
Lookingfor too many things
 Shrinkscandidatepool
 Disproportionateimpactonwomen

z
Knowledge
Should
Be ....
Hardto Acquire
--or--
A redflag tolack

Ask Yourself
Do you really need that?
Am I beingrealistic?
How long wouldit take to learn?
Do you havetheresources to train?

Best Practices
Hard to acquire ORa red flag
Relevant to job
Discussions > fact grilling
 RuleofThumb:qualitative,notquantitative,evaluation
 “Howdoes___ work?”
 “Describewhenyouwoulduse__ vs__.”

Design
Big, meaty problems

Why?
Tests:
 Abilitytotackleopen-endedproblems
 Communication/teamworkskills
 Adifferentsideofproblem-solving
Respects experience of senior candidates
47

Best Practices
Ask open-ended problems
Encourage follow up questions
Have candidate walk through
Let candidatedrive
48

Evaluation
Ability to make tradeoffs
Ability to identify issues
Separate knowledge from attributes
Response to feedback
49

Bad Practices
Unreasonable knowledgeexpectations
Some candidates don’t know how it works
 Can’taskquestions
 More ofa factualanswer
 Don’tdrivetheprocess
50

Examples
DesignAPI for…
System for Amazon book rank
System for TinyURL
OOD for a music library
51

Problem Solving
The smarts

Why?
Smart people do good work
Hires adaptable people
Very effectiveifdonewell
53

Bad questions
 Tooeasy
 Toocommon
 Toomuchofa “trick”
Bad process/coaching
Bad evaluation techniques

What makes a good question?
Medium orhard
Reasonable knowledge expectations
Not common
No single“a ha” moment

How doyou ask them effectively?
Coach candidates in helpfulways
 “Whataboutanexample?”
Give hints
Remindthem of details
Be positive and reassuring

How doyou evaluatethem?
Not about right vs. wrong
Look for great solutions
Take into account hintsand difficulty

Coding
What you need and what you don’t

Why?
Code quality matters
Not everyone can translatealgorithm into code

Evaluation
Look at structure and style
 Butdifferentiatewhat’strainable
Not about complete vs. incomplete
Let the candidate test

Bad Practices
Requiring every detail
Tedious questions
Taking over the testing
Lettingthe candidate code too early

Goal: “Seeminglycompilable” code.
Don’t waste time
 DoyoureallyneedthatNodeclass?
 Encourageabbreviations,skippinguninterestingparts,etc.
Let the candidate test (first)
Best Practices

Where do you go
from here?
Rolling out a great process

Define What YouValue
Do you have a definedculture?
Do you need specific, deep knowledge?
Whichspecificexperiences matter?
Do you solve hard-ish problems?

TrainYour Interviewers
Need dev-specific interview training
 Teachwhyyouasksomequestions
• … and why you shouldn’t teach others
 Teachthemtocreate goodexperiences
 Teachthemhowtoevaluate
Shadow & evaluate interviewers
Structure your interviews

Keep Your CandidatesHappy
Prepare your candidates
 Theyshouldknowwhattoexpect
Select friendly, effectiveinterviewers
Move quickly
Keepin touch!

No process is perfect
Find theleast-bad option
Be aware of its weaknesses
Fix as you go
THINK!

THANK YOU
gayle@gayle.com
gayle in/gaylemcdgayle

Creating the (Im)perfect Developer Interview

More Related Content

What's hot (20)

Similar to Creating the (Im)perfect Developer Interview (20)

Recently uploaded (20)

Creating the (Im)perfect Developer Interview