January 9, 2016
Introduction
Julia
spent time again in 2016 to review the video. She likes to take some notes, and
then, look into them. So, she remembers now that interview white board question
solving should be fun, and should be creative, and should be easy to work with,
since the interviewer likes to drop hints, and make you happy and see if
you can solve the problem, or learn something from you as well.
Moishe
Lettvin - What I Learned
Doing 250 Interviews at Google
Moishe was in the hiring committee, and also taught the interview
process.
Write your own notes
Watch again in 2016, and this time Julia likes to take notes about 25 things about interview.
1. It's Fun!
Talk
to new people, meeting new people is extremely fun, sharing expertise with
people who have same knowledge - great fun.
Learn through interview, how to interact, no matter undergraduate dropout or PH.D.
2. Except when it isn't
3
hours for preparation, before and after; meet the strangers
Random
things happens -
Candidate
cries, cannot figure out the solution;
Some
one just refused to write code; communication problems; People are jerks;
people are refused to write code; for more senior position etc.
3. Noisy, inaccurate, arbitrary (random, casual)
To
write code on white board, it is difficult to do
some
cascading problem
Noisy
- interviewer may have a bad day, how to feel about interacting with a person
Interviewer may be in a bad day; you can get different results
If
you are in border line, you are invited to come back in 1 year - google is just
generous
5
people can say no if they interview any hired person
Google's
response:
Set
the bar very high - false positive, false negative - not hire a bad person
For
example, one week, 8 of candidates, say no to all of them
4. Data is a blessing and a curse
no
correlation between interviewee and ...
each
interviewer give 1 - 4, 4 definitively hire, 2.0 - 3.0 range
You
can't put a number on stuff that matters
Written feedback - 3 or 4 pages narrative about the person
Same
as IQ, it is not accurate
You can't even ask about stuff that matters - things to love work with, things about employee
How
to test it in the interview?
Give
them the question, see if she/he has passion, and their creative ability, their
enthusiasm, maybe their sense of fun, pretty good habit to hack, come out the
solution
Keypoint
is like a criticism, not statistics.
5. It's a team effort, and isolation hurts
One
interviewer goes very deep in one area; two personalities clicks very
well.
Microsoft
- hiring manager is responsible for saying yes/ no - hiring bar goes down a
little bit
but
Google, Hiring for Google, not for hiring for a team
Analysis of hiring of Microsoft:
not be on malicious, just need to hire some one; bar will keep going down; no
check on it.
Analysis
of hiring of Google:
very consistent.
6. Be prepared!
When
you have to give out interview, you should be prepared.
Having
a structure will help: Google has 4-5 hiring - interview process
class - basic outline
10
minutes -
40
minutes - question
10
minutes - question to answer
More nervous than interviewee - first interview to give.
7. But, it's like jazz
unexpected
thing will happen, will go side ways;
Favorite
interview questions:
Maybe
they are heard; you can figure out quickly; improvise,
Maybe
they do not like to write code
Maybe
people have better ideas to write code
Maybe
people can
Maybe
be taught, interviewee may give out total out-of-surprise solution
8. Please make mistakes!
It
takes practice. calibrates, 3 people gives interviews together.
people
did more than 25 interviews.
9. Every interview is a conversation (with goals)
It
is not a test, or something like putting through and see if the interviewee can
survive
Conversation has goals, which is ok; Do not just stress the interviewer out;
Primary goal:
Happy
candidate
Make
them happy
Do
not get something from the candidate, but make them feel not good.
It
does not take much to gain notoriety - people blog about the bad interview
process
Interview is stressful Survey: Do you have good experience of Google interview?
Other
goal: Answer "Would I love to work with this person?"
Even
the people is not in the same team, do you like to work with him; contribute
the company, can he teach you, can they be taught by you, can he inspire you
etc. ? Idea world, work with the people smart than me. You want to keep raising
the bar.
Is
this person having qualities, and then people love to work with? Can they
contribute to the company
Other
goal: give answer to "Would you love to work at this company?"
Talk
about culture, how excited to work here.
Feel
real Humanity sense during the interview while I am working here.
Good
interviewers are generous - Ask good questions when interview goes on. Honest
to candidate, work hard on candidate. Ask good questions in the interview goes
on.
I
learn something amazing, graph traversal problem for example, traverse the
graph, and find all the words in them.
Second
part, implement a dictionary to look up those words quickly. Use tries for
that.
That
is what far the interviewer gets on that question. But he asked the candidate,
he got through that approximately 8 minutes. What else you can do, he showed me
that traverse the graph and trie in parallel, magic and beautiful thing.
cannot
write down the detail.
Example:
1.
Traverse graph, and build up a dictionary. A candidate spends 8 minutes to figure
out. The candidate shows doing thing in parallel.
2.
Random number, time stamp
superior
smart, give me direction
3.
Beauty math and engineering problem
Interviewer
being prepared, knowledge
10. Time
is of the essence - counter point of generous of interviewer
Time
management - easy to get away from you as an interviewer.
Candidate
may talk too much not important, not relate to the conversation the interviewer
likes to have.
A
few thoughts on technical interviews:
11. Good
question are like onions
naive
solution in the crust, you can peel back, deep, more interesting problems going
on.
First
technical interview question I asked, write a program to draw a circle -
Microsoft,
1994, super nervous, five minutes small talk
Write
a program to draw a circle - sin, cos, 2 * PI,
Interviewer
says that it sucks, speed up
sqrt
root
All
right, good job
-
15 minutes left, no square root - compute error - with right hint
compute
error - high far away you are
congratulations:
circle drawing algorithms by windows
questions
goes to deeper and deeper, for the candidate, it is great. More iterate on
something
Iterative
on something
12. Strive
for higher bandwidth
prefer
for the white board to just talk to people, can get more data
First
Microsoft interview, whole nervous whole day.
Just
give me a notebook and problem, and let me work on the problems.
13. Artifice
is inevitable
Being
a candidate is not having a job.
Is
this person enjoyable to be around? Is this person creative? Can this person
learn?
14. More
signal; less noise
Let
candidate know what they are here for; let interviewee prepare
Your
job as an interviewer is to help them show their best work. That is what you
want to see. You are not just trying to make them fail.
15. This
might be the best we can do
It
probably isn't
Hack
school is a good hiring tool - 2 months of coding, some one can finish or so
on.
15. Have
Fun
Learn
Questions?
1.
Machine learning helps the interviewer to decide ...
Human
input, the written feedback, hiring committee
2.
Question: false positive
Worthy
taking some risk on some one; Make a bad hire, cascading effects,
3.
Microsoft hires for a team, Google hires for the company
hiring
committee -
culture
thing -
4.
How to hire people with domain knowledge? 1 person with domain knowledge
5.
Personality fit vs strong technique skills
Ability
to lean and pick things up; demonstrate Java and Java Script knowledge, so
he/she can pick up PHP.
People
cannot get the expertise you need.
6.
Don't hire ass holes if he is perfect for the job.
During
interview, he is a domain expert; Some one creates a toxic environment.
He
is bad, drag people down;
7.
Interviewers talk about, send an email for these emails.
Microsoft
email chain - hiring or not hiring / quiet sure
Poised
or blessed by your first interviewer - a senior people if he is.
Google
does not allow to communicate, every one does not have context, what happened
before. Without bias is better.
8.
Group interview - 2 interviewer, one interviewee
2
shade interviews - 2 shade interviews, just observe, do not say a word.
9.
If you have power to change, what to change?
Could
you hang out? Could you talk? Can you communicate?
Do
you have enough data for you to decide if you should hire?
10.
calibrate? new bees? How much weight?
Good
questions, exercise the candidate
here
is the place the interviewer has to hint
Apply
something
Write
solid code
Talk
through ...
11.
Who is focusing on? before interview
After
the interview, people talk about the impressions, sort of things.
2-3
pages - take up code sample - two pages typical for feedback
If
every one goes to wrap up meeting,
12.
Google does follow up, it takes months to make a decision; now it only takes
weeks.
Stumble
early on, give hints on - just more you ask question, the more you know how to
give hints.
13.
behavior questions:
agree
general ideas how to do behavior questions.
Program
interview exposed.
14.
turnaround time - how quickly to get back to the candidate?
Microsoft
- last interview - you have the hiring manager to interview you.
Google
- April, interview, at the end of June.
Nov.
30, 2016
- Proficiently
-
Quickly <- ask questions
-
without mistakes
-
Good practices
January 14, 2017
Read quora answers provided by Moishe Lettvin.
Answer of Google interview, very good time to read the answer.
No degree, imposter syndrome; Julia did not get her Ph.D. degree, but she had a special experience on her Ph.D. study, she went through the immigration process and then became a Canadian citizen. She then went back to her career pursuit relentlessly through Hackerrank contest training.
Answer of Google interview, very good time to read the answer.
No degree, imposter syndrome; Julia did not get her Ph.D. degree, but she had a special experience on her Ph.D. study, she went through the immigration process and then became a Canadian citizen. She then went back to her career pursuit relentlessly through Hackerrank contest training.
No comments:
Post a Comment