Subscribe
Share
Share
Embed
课代表立正的官方Podcast
深度访谈，有用干货，亲身验证的「真本事」
Superlinear Academy创始人，Maven Top AI Instructor
前Statsig布道师（OpenAI收购），腾讯副总监，Meta，Amazon；康奈尔经济学博士
社区：Superlinear.Academy
课程：ai-builders.com
个人：lizheng.ai
uh so today i would like to introduce
our speaker
uh who is also a cornell uh
student but right now he after he
graduates from
cornell he actually has a lot of
[Music]
decisions on how to make a change from
one job to another
he was he had experience in building
setups
and he also had some interesting
experience in pcg
so and later after he choose to not to
do consulting anymore
and to switch to amazon and
right now he's a senior uh data
scientist on facebook
so i feel uh it's you know he has
this is his name is being tied with so
many big names
of these you know great companies and
especially
i do during our personal interaction i
also know
eugene han also wanted to stay in
academia but
later he made the transition and go to
uh you know not academia world so i'm
really excited
to be able to have him as our
second uh the speaker of our second
episode
and i i think you will also
learn a lot uh especially at this stage
of your phd
probably you are having some uh
you know you probably want to stay in
academy and you also want to uh
explore the world outside and i think
you will benefit a lot from today's talk
so be who can introduce the rules yeah
um just a couple of housekeeping rules
before we start
so we would appreciate it if you meet
yourself during a talk
to eliminate background noise but you
know feel free to open your
camera if you want um and then at the
end of the talk we'll
leave a lot of time for q a which i
think is
going to be very the most the more
interesting part where you can actually
directly ask each and any questions
if you want to ask directly um then
just type raise hand in the chat box and
then we will give the mic to you
if you're a bit shy you can type in the
questions in the chat box and then one
of us will read it out loud
please don't ask too many questions
maybe not more than two but
definitely encourage everyone to at
least ask one question
um i think oh last thing um
when you're asking questions it'd be
great to turn on your video because it
just
always feels more personal to see the
person face to face so
that's pretty much everything should we
give the floor to our speaker
definitely so i just stopped sharing you
john can you use
start sharing hey
yeah so thank you for inviting us thank
you for having this uh series
i'm going to talk about that in my
presentation as well because uh
that is one of the biggest lessons
learned i learned as well
like looking back to my phd years that i
didn't
spend much time knowing what's out there
so just last stage of my academia pass
for too long and that is one of my
biggest regret
of my of my life otherwise i'll be much
richer than
what i am right now okay so thank you
and uh one more thing
sometimes i do not control the pace of
my
pace of my speak well and uh if you
find something uh unclear uh feel free
to interrupt me and
ask me to repeat uh what do i actually
mean
all right i'll start sharing
okay so from e-com phd to
tag because the first part is going to
be
a quick introduction of my journey and
the second is my lessons and uh kind of
my
three advices so self intro this is just
a screenshot of my linking
uh right now i'm a synergist center it's
on facebook i do
facebook e-commerce uh not the facebook
shops
the marketplace citizenship so if you
i guess most of us uh are from china
it's kind of shiny
in the american version and uh before
that i was an economist at
amazon so we do so that okay
so the difference between these two jobs
uh at amazon as an economist
i get to do a lot of uh what i do in phd
so i read academic papers i do models i
run machine learning causal inference
models
but as a data scientist at facebook
i mostly just do sql
to get data and then tableau to plot two
charts basically just plot bar charts
and then i do kind of analytics
so i tell people what does the data mean
and
how does that drives our decisions uh
so i'm basically using i guess second
year
statistical information i'm not doing
anything
that is related to my phd education and
uh
i'm actually pretty happy about my
current job and
i'll talk later about uh past dependent
and uh you wish we don't need to uh
just use what we learned we should uh
and constrain optimization
but let's save the talk later and before
that
uh it was cornell phd in economics
and uh it's six years and i hope this
stays as four years that i'll have i
would have much better
much more experience and join tech
companies much earlier
but okay on the right is my dissertation
so the
dissertation is about uh firms investors
and stock markets
uh i my my pitch my
job market paper is the real impact of
ipl so using some natural shock
happening china to understand the real
impact of ipo on
like the impact of ipo on real economy
but now when i look back into my
dissociation i don't even care about
what conclusion i drew because it has no
impact on the real world
okay and before i graduate so i
graduated
in um august
2017 and before i graduate i spent
three months uh as a summer consultant
at bcg
in beijing office i liked the experience
a lot but for some personal
reasons i chose to um came back to
the united states and uh do uh
being tech and during my phd
so these are during my phd i had some
very uh
strange i guess
abnormal or not normal experiences
as a phd student we founded
ramen shop in new york city and we found
it we did some
i guess college grocery delivery uh
business
at cornell so it's uh almost like amazon
fresh
we we run weekly delivery from new york
city
to cornell campuses and delivered to
student dorms
so definitely eyebrow uh these are some
screenshots
this is the ramen shop and we did a
holiday market by a korean town
that was like we go there every morning
and
uh the market starts at 11 i think
then we do we just make japanese ramen
and uh we uh close at i
i forgot maybe 9 p.m and then we wash
the pots the dishes and do it all over
again the next day
not not me so but yeah i helped
and this this is the picture for the
college grocery
so we made posters uh we collaborate
with a
new york city chinese grocery store and
we deliver other stuff
to campuses so this is you you get to
order on our website
we get your stuff in your city and we
deliver to you and this is kind of our
delivery so this car uh goes from new
york city to
ithaca and in up in boxes so a hardcore
share that new york city
puts the stuff into boxes and they will
sort into individual orders
we had to do this because uh we didn't
have a big enough car
from new york city to ithaca
all right so all those
experiences right uh what
what what is my biggest lesson the
biggest lesson is i enjoy my life much
better than
after phd than i was during phd
so at doing phd i i also lived uh
i don't know like i like to explore i
like to do different things
but one i have this constant fear of
feeling that what if i don't get my phd
what if i fail
at academia my peers are studying and
doing research
and that's so much smarter than me uh
what i'm doing
you know enjoying my life so i like if i
go out
and uh just have some fun i would have
a lot of self-guilt and that is not a
very pleasant feeling
uh if you when you are you anxiety
and fear i guess so uh
after phd because i sell my time to the
company
i get to separate work versus life no
matter how much
hours i put in my work i get to enjoy my
own life
so an anecdote story is uh
in pcg beijing uh the hours were quite
uh crazy so i had to work uh nine a.m
to about uh average two a.m
sometimes i go home at 12 and i'll feel
like
today is a very lucky day because i can
i get to go home
uh 12 10 a.m and sometimes you have to
go home
at 5 00 yeah and wake up the next day we
kept the same day
sleeping like three hours but i still
felt better
than my phd that was during my phd time
like that was before my graduation
i still feel better uh doing the
consulting work uh with such a long hour
than what had done my experience during
phd because
uh i got to separate work and life so
and i get to quite like be myself
basically
after work and that feels very good and
the last thing to mention is the living
expenses you get to earn much more
after phd so that's a big bonus
and uh this is kind of my instagram
story archive i just
so not happening since i graduated from
phd
i got to do things yeah this lesson
number one
and if i could do my phd all over again
uh what would i do so first
i didn't have this uh theory i didn't
have this
world beyond academia theory when i was
in phd
if i had that maybe i can realize much
sooner
that academia is not what i want to do
so
the first thing i want to do is admit
that i don't want to do academia in my
second year instead of
my sixth year i stayed in phd for six
years
because i always think that you know
the best people in my best peers that
are doing academia
and that is how the how phd is supposed
to
how how we are supposed to do academia
so
i came here to do academia i should do
academia
if i feel like i don't want to do
leukemia that is just because
things things are getting too hard so
i'm quitting
i equate uh not interested in academia
as funding excuse for quitting so i just
stayed
for too long if i realized i don't
like if i realized that honestly i do
not want to do phd
i would have told my committee that i
don't want to do academia as soon as i
found that
and then i'll start looking for industry
internships
so i didn't do any industry internships
during my phd
so the bcg is my first kind of
internship
experience during my phd i would have
done in industry internships
and that would prepare me for much
better
position or a situation
when i was looking for industry jobs
then i would have finished my phd in
four years instead of six
because i don't think there was any
extra value for me of
the society by me staying in my phd
program for six years
uh and uh one thing i'm not sure if it's
achievable but maybe if i realize i
don't want to do academia
that is during my phd i should have
separated my work
from my life uh i should have just had
work hours and
not think about the research and the
other stuff i do after the work hours
because
when i think back uh i have
self guilt i feel guilty uh
not working so i constantly think about
my work
but it didn't help my productivity i
feel like
looking back if i separated my work and
the life i could be more productive in
work and must be
much happier in life so that is uh
another lesson
all right so three advices after having
this
lesson and uh spending a couple years
after academia
advice number two is uh know yourself
because i didn't
and i wasn't completely honest with
myself
and i think it's a very easy bias like
like not knowing yourself is common
knowing yourself is difficult
so know what is your true interest and
your strengths
like we are all good at studying well
good at reading papers
i'm not good at doing research that's
why we are here
but uh beyond that what is your
strengths and
most most importantly what actually
drives you
like do you feel excited doing research
uh so i was okay so this part the reason
part of the reason i stayed in academia
for very long is i i was actually very
interested in research
i was in empirical corporate finance and
that actually found
uh the papers very interesting and i
enjoy reading papers and i enjoy doing
research
but even even with that
i still feel like i do not want to do
academia because what actually drives me
is having an impact uh
on the world i don't feel like i was
doing that
uh by doing research like the paper i
write
i don't think it's it can actually
create
uh the impact i wanted to i want to i
want for myself
um yeah so
understanding your underlying motivation
is very important
and then have a com like be honest about
what are your resources
and what you can do and sometimes
knowing this
takes trying like you don't know what
you don't know right
so you have to try uh like intact they
have to build this
meaningful minimum allowable products to
rather detail assumptions
uh sometimes it doesn't take much for
example coming to
this series and listening to was beyond
academia
is a way of knowing uh doing
internships is a great way to try
if you like to really get to know the
day-to-day of
different jobs and uh really uh
understand
if that is what you want to do right and
also when you try now there are a lot of
two-way door decisions especially when
you are
in phd so for example if you do an
internship in
industry if you don't like it you can
still come back
you can still do academia so a lot of
decisions are two-way door that means
you can
just try and come back there is no not
much harm for trying
so uh definitely take those
opportunities
and uh try like the industry the
industry internship
or the industry experiences if even if
you decide to do academia
would be good for your uh research and
for academic job because you know
like for example what kind of data is
out there what people actually care
about
and maybe you build some connections
maybe like the
like having multiple dimension of
experiences
is going to help you not going to hurt
you
otherwise number two is unconstrained
optimization
is always better than constrained number
maybe a
bigger equal to bigger than or equal to
but it's
like most of the case is better right
sometimes it's equal
um now sometimes we're stuck in female
for too long and
that is i guess echoes uh because
earlier points
that a lot of people stay in academia
because one they don't know what they
can do
outside of academia and two uh
they want to use they spend so much time
in academia they spend so much time in
research they want to use that
they want to play that strength and uh
they don't want to just give it up
but that is uh i guess bias itself
like just think about what makes you
happy tomorrow
or in 10 years or in 20 years of your
life
it's not going to be like uh a lot of
cases is not going to be
just you know repeating what you did for
the past 10 years
like if today there are two choices
option number one
leads to a better life in 10 years or
even
in one year option number two is doing
what you
have been doing or what you are good at
uh but
you are going to be less happy right
here choose option number one don't
don't choose option number two just
because you did
option number two for such a long time
right
so do unconstrained optimization don't
let your past
uh constrain you so past dependent
sound caused lots of version those are
things that stop you from trying new
things
and uh i talked about this earlier like
really be honest about yourself not only
like
admitting uh your strengths and your
shortcomings
but also you know sometimes it's so
okay to quit and know that it's okay to
quit
i was i guess for me it was my ego
and pride or like my self uh
discipline that got into my way of
admitting i'm not interested in academia
so it goes both ways right so rather be
completely honest with yourself and the
cost
the last the last the the second part is
the cost a lot of times the cost of not
not acting
is bigger than the risk of acting and
phd
the program definitely gives you a
shelter an excuse to not act
what do i mean by this if you are a phd
investing schools right it's five years
program it is
it is a very long time and and
the program doesn't ask you much results
until the end because they want to give
you some freedom
to uh to really focus and do
research without the worry that
you know next week out what i have to do
next month what i have to do
they give you some stipends and they
give you a very
uh lenient environment and to tell you
you know just focus on research
so it's a shelter to not trying
different things because
for example for me before every summer i
stayed at cornell
and uh did extra research i you already
applied for a summer ra or ta
and then i stayed at school and i go to
library every day to do research
instead of applying for an internship
and i felt pretty good about my decision
because i was focusing on research i
didn't
waste any of my time but uh
now if you ask me what did i do
summer of uh i don't know 2015
i don't remember i don't remember what
paper i read i don't remember what
research i did
i feel like it's so completely it's not
complete waste of time
but it is a waste of time so uh
that was the cost of me not having uh
not acting to apply for a summer
internship
and uh i guess that's an easy trap
for a lot of phd because the program is
set up this way
so think about what is the cost of you
not acting
and uh what is the risk if you actually
act a lot of the time the risk of acting
is very small
you actually benefit from acting and the
cause of not acting
is uh ignored advice number three
is uh no really know what's out there
and this one i think is the
i guess in terms of risk reward it's
definitely
everyone should just know uh like you
know
you should know your options beyond
academia right so coming to this talk
again is great
the options uh what industries
you can go like i hear a lot of you
either come from consulting
or come from uh i want to do consulting
consulting is a great industry
i later i'll like post a screenshot of
my
youtube channel and
i talked about the value of consulting
uh that brings to the to to the firm
that hires consultant
i think it's a great industry but beyond
consulting
there are also a lot of other uh
industry
industries that just have a phd because
you are a phd
for example tech hires all sorts of
to solve their problems whether it's
data
whether it's optimization uh like some
uh
operation research and sometimes it's
even like
material science mechanical engineering
and they have all sorts of phd so know
our options
i guess consulting and data other phds
can just get into and try so
no was really like know your average
salary in our career paths
and uh what to do to day to day yeah
really know what do they do so then you
can
fairly access are you going to be
interested in this type of job
or should i do something else and also
now option if you choose to do something
else
for example if you just take a computer
science master
or if you go to go do some
internship or if you prepare an
interview
for half a year or maybe you take some
i don't know machine learning courses
and you build your resume that way
it will open a lot of different options
for you so know uh based on
your current capacity and uh based on
what you can do
uh in a year no options and really
do the unconstrained optimization and
then the i guess the last part is chance
of success
and that leads to my next slide
how to increase our chance of success of
getting what you want to
uh what you try to get so
fence the gap between what you want to
do the first is
know what you want to do right and no
other industries
know what like what career you want to
choose
for example a lot of you want to do
consulting that's great then what is the
gap between your current resume
between current knowledge experiences
projects to have the right internships
to get to consulting are they going to
take your resume and give you an
interview
and then you need to prepare for
interviews right and
sometimes building connections can give
you
either recommendations referral or just
give you some knowledge and
you know tell you
what's the most effective way to prepare
for interview or
to bridge our gap so this is
i guess an action item know what's out
there
have access like a fair honest access
access process of yourself
and then start acting to close the gap
i have a youtube video about data
science what data scientist
data scientists do and today i'm going
to
uh meet with a friend and shoot another
video to talk about
uh how to close the gap to get into data
scientists
data science so what companies are
looking for
and uh how how do you prepare what
materials do you need to do
to prepare for to close the gap so stay
tuned and
subscribe all right so this is a
screenshot of my
youtube channel uh for example talk
about ap testing talk about how ai is
applied
in big tech companies what is management
consulting
and what is data science i have some
jojo articles
maybe in the future about uh also making
videos based on these articles
but this one for example talks about how
you prepare for mock interview
um i guess okay a little bit about
myself
maybe here yeah it's all going to be
okay yeah a little bit about myself
in the end so one uh because you are a
phd
you are all smart and diligent otherwise
you won't be able to
get here um people a lot of people like
a lot of us
have uh impulsive syndrome uh that you
know
we got here because of uh lag or because
uh but everyone all my peers are smarter
than me so i must be the
imposer i must be faking it no you
deserve
the phd you're smart intelligent
and you actually have better resources
than most people i'm not only talking
about
the your advisors your professors
the class you can take you can take most
classes for free as a phd
so uh take those classes do the projects
build your resume
but also like uh even if you try to fake
right if you want to
for example people a lot of people in in
tag
if like ai is is very easy
to fake not too fake but very easy to
uh to to get to know uh for example
machine learning is just uh
basically it takes uh takes a month or
or or maybe three months
to to get to to know ai well
if you already prepared for it so if you
take three months
and prepare like learn about ai and then
you take another three months write a
paper about ai
uh maybe it's not it's useful maybe it's
not
but hr because they are phd will pay
attention to you
and it will add uh it will boost your
resume
and give you a chance to most tech
companies
so that is the resource that many people
do not have
uh then you can maximize your experience
on random you maximize the
experience on one dimension um so i
when i was doing consulting i was back
in beijing
and my i guess my
peers people who are at my same age i
held with people who do
startups and a lot of them
were quite successful they started
companies
that are well-known names they they had
like if they went to industry uh some of
them are directors
they were quite successful but i feel
like i didn't waste my time
doing phd because i met i also like
because i had type of phd
i maximized my experience on academia
and
uh so ph doing doing phd
in academia is really like playing at a
very high level
right and not many people get to see
that so if
you draw an analogy in sports it's
almost like playing a youth team for
football club right you got to the top
level you know
what the top level looks like or like if
you found it a
mildly successful successful startup i
feel like you don't know what
what does it take to be aware to be a
true uh
truly successful startup but you like
you interact
with the best scholars in the world and
uh you have this title
that stays with you for your whole life
that says you know you're a phd
you complete uh a great deal in academia
and you are one of the smartest people
uh in the world
so that's great it's all going to be
okay um
i want to oh one last thing about myself
yeah uh i wasted
not wasted i spent six years in academia
and only to realize all my six
years but i don't really want to do
academia
so i applied the bcg uh going back to my
resume
i played i applied the bcg uh
job i think december
2016 and
got to the internship january 27th
oh because i got the interview january
2017.
i prepared a week for my mock interviews
and got the internship and got in
so one is a head article about how to
prepare mock interviews
how to prepare the interviews
efficiently but two
i feel like that's a lucky shot
and then at amazon like i didn't apply
to
any tech jobs amazon is the only one
applied i got the interview and i got in
and later i discovered that is kind of
one of the
best jobs uh that i can i can ever have
so a lot of this is luck but
i don't think it's purely luck it's also
because i have the title
i had some experiences during phd
and i actually spent the time and
already prepared for the interview
so there are a lot of things you can do
much better than i
i did by starting early so that was all
my advice and lessons
if i knew this in my second year
i would prepare myself so well that i
couldn't fail
so yeah that is my lessons and devices
thank you for listening
thank you aujen that's very helpful and
i think
you know what's really impressive is how
genuine you are
in sharing your thoughts and experience
and it's really encouraging to people
who
might be having second doubts about
themselves um
you know before people in the audience
feel
free to ask questions or type in your
questions
but we've had a couple coming in and i
feel like
because your experience is so broad
people might ask all sorts of
um questions and it could be you know
loosely
categorized into your experience as an
economist
and then you know having startups
yourselves
and then transitioning into a data
scientist
and now you know having your own youtube
channel
and then some people might just want
some general advice
for you so let's try to maybe group our
questions in this kind of categories
um anyone wants to be the first person
asking a question
live
hello hello
hi fumi we can hear you okay uh
so would you advise to uh
someone who is interested in going to
the industry later
sorry for too enable of the video
sharing because that's the rule we set
for q a if you want to ask questions
what's the rule yeah there's a rule for
q a if you want to ask
in person it's better to just share a
video
oh to turn on the video correct
oh uh unnecessary
all right sorry i didn't prepare
okay yeah feel free okay
uh would it be better to try to finish
the phd faster
or utilize the summer to do internships
and how would you compare the problems
in financial markets
to consumer markets two questions
yes well the first one i think is easy
but i didn't get the second one to ask
for clarification the first one like uh
internship is really to try out right
and build a resume
if you have a like good enough resume
and if you know what you need to do then
just finish your phd and do that
but if you don't then definitely you
know
try as many internships and until you
you have a good enough resume to do the
things you know
you want to do uh the second question
financial markets and the consumer
markets
was that like is it choosing between the
industry or
are the problems similar or different
the promise
are the problems similar or different
oh uh okay i guess it's about
how did i apply my my knowledge of
a finance phd like economics
concentrated in finance
to consumer markets right um a lot of
tools and
skills that can be carried over for
example i did
apply economic econometrics i did cause
inference for the corporate finance
study for example matching difference in
difference
all those causal inference stuff and
those twos
can be directly applied to solve
consumer markets problems that's why
amazon hired me
to do or what they hired me to do
but uh again going back to not limiting
myself
i guess i didn't uh this reminds me i
didn't i promised to talk about the
difference between data scientists and
economics and uh forgot to do that uh
like it doesn't like what drives me
is to solve problems and uh crazy impact
and it doesn't take complicated model to
solve problems
uh in a lot of cases if the problem is
worth stop solving
like comparing the two means or you know
finding where the problem is
is much more important than being able
to build sophisticated models
of course there are problems uh that
needs to be solved by sophisticated
models and uh
i'm also like mildly interested in doing
that
but uh that that is not um
that is not how i think about my career
when i think about my career is not
about how to apply
my past knowledge it's about what
problems excites me
so i guess to some to sum up some of the
skills
are transferable but the domain
knowledge for example is definitely not
relevant
and yeah i don't feel sorry for that
right thank you anyone else got a
question
uh can everyone help me yes
uh hi i'm martin i have one question
about uh
the stop so i'm curious what happened to
your stop
in the end oh we closed
college grocery well with the car
because it's actually a
very heavy operation like it takes a lot
of operations for example
you have to do customer service you have
to maintain a good relationship with the
grocery store
uh if you get uh one kind of uh vinegar
but they're out of stock do you replace
or do you
refund there are a lot of uh like tricky
uh
like things that takes time and nobody
was willing to do that after i left
so just
operating in new york city i hope the
pandemic didn't
didn't make them bankrupt but uh they
had
three i think
shops in new york city
okay i think yeah so i have a second
question which is
a little personal so what's the uh
intensive when you
decided to uh uh like build up a
stop is it like to try to make
a lot of money or like just a random
decision to try different things
yeah a good question because that's
actually my motivation for doing a lot
of things including starter youtube
channel it's not
it's not it's never about the money it's
just about
seeing a problem and i want to solve it
and i have a huge like a big drive to
solve the problem
uh the start started because i go to new
york city quite often
and every time i go there people ask me
can you bring us
some groceries bag i'm willing to pay
you like 50
uh above the recipe uh like uh what
like the the market price and i said
uh okay there is a real demand uh let's
try to solve it
and uh that's that's how it started we
started with a google form
and then turned into a week's website
and then turned into a shopify
website uh in ritual respect
it actually helped a lot of my current
uh experiences
because facebook is doing the facebook
shops with shopify
and my experiences of running a shopify
store
actually helped me understand a lot of
e-commerce business in general
but uh that was never the plan
i said thank you
uh yeah let me ask you one question
okay uh thank you regen uh i'm also a
phd student in economics so
i know that amazon has already
established the economics team
while facebook is rather new in building
its uh economics team so
i'm curious that how do you feel about
working in these two companies
uh do you feel more like familiar in
amazon environment compared to in
facebook
there are two very different companies
even though they are intact i would say
there is a more difference between
facebook and amazon and between amazon
and
bcg even uh and i don't think facebook
is going to hire a lot of
economists to do uh the economist
because at facebook it provides a pro
free product
they can run a b testing for almost
everything and amazon
hires economics mostly to do causal
inference
but facebook can do a b testing they
don't they don't need to do this
but my training of economics phd
helped me to uh i guess the training is
helpful
uh i also have a very short video about
this it forced me to think about
equilibrium
and think about marginal and when you
solve problems like tech companies
um your friends a lot of people they
focus on the first order
effect of your policy for example if you
set up press 4 right
it will make people sell more expensive
stuff
but they didn't think about the
substitution effect or the incoming
factor or whatever they don't care about
they didn't think about the second order
problem
and a lot of a lot of the time the
second uh
the second order impact is more
important so uh
i guess the training of economics phd
if you have the sufficient tech skill
like if you can write sql well if you
can do analysis fast
if you can do data manipulation in
python
then combined with the economics degree
is quite welcomed at any company
consulting or tag
but yeah
like uh don't expect other tech
companies to hire
economists as amazon does yeah
thank you i have a related question on
that actually
um when i was when i first joined
mckinsey i actually spent three months
working on the data science project
and my role was what they call a data
translator so
our project was around um something
that's got to do with clinical and
pharmaceutical applications of data
science and we had
half of our team pure data scientists
and data engineers and the other have
people
with a medical degree or phd in
biological sciences
so i'm wondering that you know outside
of maybe you know consulting
related data science projects in the
real data science industry
is there a need for this kind of roles
of people who
may not be coding but they know
something about data science
to translate the business or clinical or
consumer aspect of
the industry
yes and but uh okay
so the the world of consulting is
pretty different than a word of tag uh
and related to a question in this way
consulting deals with a lot of companies
right they don't necessarily have the
expertise
to understand those complex analysis or
complex research
so they need a translator
in tech everyone should know tech like
uh they have a very high hiring bar
for uh ba engineers
or product manager so everybody knows
their shift basically
they don't need a translator to
translate otherwise they'll
just fail at their job because that's
the expectation they know their stuff
um that being said said uh like
for example people in business
function they don't necessarily uh
that they're not they don't they're not
necessarily well-versed in
uh in models for example they understand
what's the difference between
correlation and causation they know to
get this uh
you can't you have to overcome selection
bias you know
you know on a very high level what
machine learning can do
and cannot cannot do but sometimes if
you want to for example translate a
device to mean
and uh like a normal mean how does that
translate into business impact
they don't know that they don't they
don't have that uh knowledge uh
for granted so
i guess in order to be successful
in tech as a phd knowing
how have the domain expertise and have
the
knowledge the skill is important but
being able to communicate
to a way that most people like engineers
and
pms you can expect them to have a
certain type of
level of knowledge but not as sophistic
as sophisticated as you are
um being able to communicate is also
important
so that is the skill
that is not the type of job that is more
of a skill that you
you should have in order to be
successful to be successful
um i have a question sorry
uh go ahead as uh thanks for uh
each very uh informational talk so my
question is that
as a phd of economics you have all the
backgrounds of the economics and
aesthetics
so um the job for a consultant is kind
of like
fit for you because you have all the
knowledges
so um as a you know
engineering student as me i don't have a
lot of you know economics background so
do you have any advice for me to you
know get some resources
um to get those knowledges to prepare
for the
you know future consultant or you know
potential jobs at
uh these kind of fields uh good question
so the my friend who prepared me for my
interview
is a bio biochemistry phd
the first consultant i work with is a
mechanical engineering phd
i would say their last social science
phd
at consultant then there are
engineering or science phds
so and when i did the consulting
interview they didn't care about
my economics background to care about
uh how i performed in case interviews to
care
about how my business ends but they
didn't care about my degree
and a lot of times i think having a
sense a hard science
uh background uh gives you the extra
edge
because uh you can actually understand
the hard problems that your clients are
facing
so for example if your client is a
auto manufacturer then being a
mechanical engineering phd
is a very good uh boost
um so thanks so is there a fundamental
like
requirement to to you know let
the hr think oh you are at least
uh being able to understand those
economic
fundamentals is there a class i need to
take or you advise me to
take for the leadership is the best way
to signal
that you are qualified to hr so get a
pta
get an internship at a second tier
consulting company like get up
get the best internship you can get
that's possible
to hr i see thanks
i guess i can follow up on this question
so for me
personally i'm considering a transfer
future goal
is to start a startup but there are
basically
a lot of uh a lot of ways i can take to
achieve that goal
so uh right now i'm thinking about two
sense
uh the first way is to enter consulting
for a couple of years
build up my connections and then use
that connections and resources to
help me to build successful stuff a
second
but this this actually one of the sense
in the previous talk the first speaker
was a
really uh he doesn't like this idea
because he's you know running startups
and he would recommend
that instead of going to consulting i
should consider
going into joining uh joining a startup
directly
so how do you compare these two
different chats if my future goal
is to build a company
so i think that there are too many
variables in that decision
like for example if you know the startup
is going to be successful then do that
but
i guess that's uncertainty right so the
safe path is to join
uh industry build our experience and uh
you know
uh take the chance when you feel you
need to
or it's safe for you to or um
when it is its likelihood of success is
uh like
high enough but uh i guess different
people have just
different risk appetite and different
expertise
so i i i don't know how i can give a
recommendation
okay
excuse me
uh i'm curious about the like uh you're
as a
senior data scientist i'm curious about
what the most
important techniques or tools you are
using in your daily work
i wonder if there's any gaps between the
just for the data science is there a gap
between the industry and the
lab because i'm working on something
like a very complicated
deep learning or machine learning models
i am i'm
uh i'm curious if that is you useful in
your
real work industry yes yes
very very so uh i guess my two different
jobs right
facebook what i do is i run sql to get
data
and i use tableau to plot the data and
then i use ppt to present the data
that's how message and there i write
docs i guess the writing
training i received a phd was actually
quite useful
i'm able to produce high quality uh
notes
it's almost like a mini research paper
at a good speed so that helped me to
land a lot of
impact i sometimes use python to do
data cleaning and do some simple
analysis models
but i often use tableau because the
problem just requires tableau to solve
maybe testing is a must at facebook
being able to do causal inference based
on eb testing design the experiment
also data are not contaminated all those
things
are amassed at amazon
i use spark python
mostly because um the infrastructure
amazon is not as good as facebook so how
to get the data i have to
basically set up the machines the
servers to get the data
and uh build the data pipeline and clean
the data
and then use python to to uh run some
what i did is uh synthetic cohort and
measures completion so simple
simple machine learning models not deep
learning models
to draw causal inference based on big
data sets
but i feel like if you can code a deep
learning model
all the tech requirements should be easy
it's just about
preparing for the interview you know
shopping on communication
and we have a video of incoming to talk
about that
yeah thank you uh hi
can you hear me
okay uh so thank you for your
presentation and
your insight so far and uh
i'm currently a master student and my
question
is about getting a job
in the u.s after after having a phd
degree
and so as far as i know in some
countries
uh such as in germany uh sometimes uh
getting a phd degree
is not a good thing because uh sometimes
uh
sometimes uh the companies uh thinks
that uh
you are overqualified for the job and uh
sometimes they don't want to to hire you
because they will need to pay you
more uh as compared to a
a master or a bachelor degree holder
and uh so uh sometimes you know if
having a phd degree makes you less
competitive for the job
so uh my question is uh can you tell me
uh
the situation in the u.s uh i mean uh
uh can you give me some insight of like
uh
uh uh having a uh um
is like uh is uh getting a phd degree
uh is a good thing and it helps you to
to
easier uh finding a job in the u.s
yeah thank you it's my question yeah so
that's my
my first advice is not to answer a
question is to
don't get a phd if uh if your purpose is
just to
boost our job competitive like if if if
your purpose is to
make it easier to apply for a job uh
don't get
don't get into phd phd is designed for
uh
research for academia like we
happen to be trained well to to be able
to do other things but
if you don't like research or academia
it's very
difficult to survive like i found it
difficult to survive myself
even though i liked every piece of my
research and the reading papers i was
able to
like just study like do research like
for months
without doing anything else but uh even
that was not a pleasant experience
so it's not worth it to you know uh to
spend five years in phd just to be able
to increase the likelihood of getting a
job
i feel like uh if you applied to an
internship and you know
climb up the corporate letter or
whatever it's much easier
no but uh yeah and
you wanna i guess the relative com
competition or like is it easier for phd
to get a job so i think yes
because most of my phd friends
if they want to go to industry if they
want they want to find a job they can
always
do that i haven't seen one case that
they want to find a job but they
couldn't either like even if they
couldn't stay in united states they can
easily find a job uh
back home yeah so uh one more quick
question is that uh do you think that
the ranking of the university is a
a critical point i mean an important
part in in
in in finding a job i mean dude hr
people will
will pay a lot of attention to the
ranking
yes okay industry
thank you oh hey uh i have a question uh
so so you you just said that like you
are actually
um very interested in academia and very
used to the lifestyle there but uh
but it's still difficult to survive i
wonder like is that that difficult i
mean
like because like you you you graduate
from uh from cornell
and i feel like if you want to go go
into the academia you can do that right
i mean
so i'm just and also i feel like
actually like like from your description
of your past phd life i feel like
yeah i feel like you you're like it's
very suitable like
like i just wonder why you change the
path like
yeah thanks
uh if you want to be successful in
academia
it's not impossible to do uh i guess
it's not
it's not very hard to do as well or like
if you want to survive it's not hard to
do if you want to do well it's difficult
but it's still possible
uh my choice was more related to
uh that you know i just i realized that
academia is not my passion i'm
passionate about a very small part of
it the part of you know
into being intellectually challenged
solve the questions nobody
uh have stopped before like those parts
but i didn't like you know
uh rewriting the paper a hundred times
going through after with
a referee process you know discussing
all the
like details that nobody cared about in
seminars or
with their peers all those things were
like
very painful to me and that was not my
passion i feel like
i can like for example consulting is a
great place to find
the same intellectual challenge but self
problem that people actually care about
i say thank you
right any other questions from the
audience
um i think i see that there is a debate
in the chat box about the requirement
for advanced degree
in the necessity of it if you were to go
to industry i think the clarification
that we need to make here is if you
think about different career options on
the spectrum
and you know you rank it whether it's
further away from
repure research or closer and i would
probably put
industry jobs as on the closer end of
the spectrum to your actual research and
maybe
consulting banking or bc on the other
spectrum
and and there's another um dimension
which is also the requirement of
specific knowledge and expertise and
that's kind of the reverse
to the spectrum so maybe um
you probably don't 100 need a phd
to go to industry but you probably need
a master's or related
you know undergraduate degree and to be
honest
i would say higher percentage would
require a related advanced degree
but if you go to the other spectrum if
you were to go into banking consulting
or anything else it's completely not a
requirement
at all in that case they will probably
put more weight on
other factors on your cv such as your
you know the ranking of your university
um the diversity of your extracurricular
and any other passion you've shown for
their
industry or sector
yeah there's a good summary like i guess
the wet end of the
spectrum is being a professor right you
100 percent need a phd
and being a researcher probably a phd
will help
but i guess the previous question is
about you know
if i want to apply for a job uh is phd
going to
help me make be more successful if you
want to be a professor
or a researcher you wouldn't ask this
question because you know
you need a phd so it's also i guess the
my answer was to the case that the phd
is an option but not only necessity
though how should you choose and uh my
advice is don't do it
it's not worth it
um i have a question i have one more
question on a data science
data science career of yours because a
lot of people around me
seem to be wanting to be a data science
actually everyone wants to be a data
scientist nowadays and a lot of them
have zero background in anything related
they're probably
in bioinformatics or in physics in
physics or even
chemistry so i wanted to know from
people around you who are proper data
scientists
what is their background how many people
are actually from non-computer science
non-math non-economic economics
background and
and and you know what do you see our
kind of
subjects or or or you know topics or
kind of their phd areas that
could be a higher likelihood for them to
enter
data science so first data centers is
not a height bar
if a phd have done quantitative research
whether it's psychology phd or
a physics phd their skill
skill set is good enough for data
scientists
because data centers is about
application
and about solving problems analytics
application solving problems
it's not about building the the most
sophisticated models right
and they actually like companies hire
specifically for machine learning
engineers
research centers or apply centers that
you know
go and build models and solve the
cutting edge artificial
problems in artificial intelligence for
example how to make
amazon's recommendation system better
that requires
specific domain knowledge but for uh
for my from my current role data
scientist uh
as i said like second second year
second uh like sophomore
suffer more level or statistic uh
in college not in phd it's enough to
handle most
cases so if you have done quality
quantitative research
uh uh you have the skill set
uh yeah like forgot what
the question was did that answer the
question
yes yeah like most phds they're
they're qualified it's just about
the experience the and the interview
to get your resume to pass the hr stage
and then prepare the interview so the
hiring manager hires you
thank you okay
like i might have a question like you
just and
uh you mentioned that uh it would be
much better if you could have
uh start to start your job prepare
start preparing for your job like
searching at the second year of your
phd but my question is that like
usually like a lot of research takes a
long time like
much more than two years to
even get on track so but how like uh
what was your suggestion on like
determining like uh what's the starting
point like uh
because i discussed with some of my
friends on like
when to decide it is the time to quit or
to continue
like my idea is more like maybe this is
more about the rewarding you have like
the
more rewarding your rewards you have
during your phd as a
academic research and then probably will
continue more on this trend i guess at
this past um
like uh i just want to hear about like
your
uh thoughts on when to you know
decide on the on your past
yeah thank you yeah um
so i guess i need to clarify that
was my own lesson and what i would have
done differently
if i started again everyone's situation
is different
from my advice is know their interest
know your uh strengths and know yourself
basically
and then make the decision based on our
own scenario
not mine right and but for me the reason
i said i wanted to do
uh to start in my second year is
of course like my research would take a
long time to build
but uh i stayed six years i wanted to do
academia i had the country fracture
right
i wanted to do academia i i stayed those
years but i wasn't particularly proud of
my
uh research paper like it is an
important topic
i think it's a nicely written paper but
uh there was no impact and uh
because what drives me is impact right
so i i don't care
like if i if i stay in my phd for four
years
and had a much crappy paper much
crappier paper
i would actually be happier today so
that's why i i said if i do it again i
would uh
start in my second year but i know
everyone is different so
definitely just uh make a decision based
on our own scenario
yeah thank you yeah thanks
i guess to follow up on that uh you join
if there's a moment
in your current life you think back you
say okay if
i were to continue academia maybe you
know i will be a professor uh uh
do you have no so you have no moments of
regret right
just first it's not helpful to regret
things right like like i said i would do
a job
i would do it differently but i uh uh
that is more uh as advice for
like those people who are in a phd
program
uh if i choose my life i wouldn't
necessarily want to do it all over again
because
i got i'm happy with my current life
right so
everything that leads to my current life
i i think you know
that they're good uh i don't have any
particular regrets
and uh it took me six years to realize i
wasn't in academia
maybe that's uh that's what i need uh if
i
change my decision in my second year
maybe i will always have
self-doubt you know what then maybe i'll
have questions
like you just mentioned you know what
what if today i don't have what if
because i suffered those
six years and uh i realized you know
that's not my
passion okay that's a good point
and all of the jobs you did uh i know
you had experience developing
some experience in consulting and now
data scientists
what is the job you like most or you you
the bet all
you're saying the best is yet to come so
what you know what i'm thinking right
now
on my current job oh
otherwise i would have uh like i so
i guess uh some some data points to back
my claim
amazon has a very particular stock
investing uh schedule so when you go to
a tech company they give you stocks
for four years but amazon the first year
it invests
five percent second year 15 third and
fourth uh
40 each and when i joined amazon amazon
stock was at
800 something now i said uh 3000.
so i was actually underpaid uh in
facebook
when i took the facebook job and so i
got paid less
at facebook than i was at amazon but i
wanted to try the job
so i came like if i'm not happy at
facebook i would go to another job
write any other questions from the
audience
i guess we have a lot of comments uh in
the chat box
i try to speak because i i cannot
yeah yeah i was muted i have two buttons
need two
our mutes so i mute one and the other
still moved it so you cannot hear me
can you hear me now okay
so i put some uh text comments because
previously i got some technical problems
to unmute both
so it's it's it's good it's good
communication we can't talk
more so so today like i put in comments
things are changing like 20 30 years ago
if a phd the industry labs
will value you more but today can be
different
yeah i'm looking at the discussion in
the
chats i think there are two things
i want to address one is
like still like we focus on the job that
requires a phd
so if you are a phd you know that is our
advantage because these jobs
uh you all have a comparative advantage
of being a phd
so as a phd definitely focus on those
jobs or not focus but uh you know pay
more attention to those jobs
and apply and you'll be successful uh
i don't know the future but uh at least
you know they're good jobs
but the second is uh like like for
example
a lot of big companies have uh
do research they have their research
department
yes and uh but don't take
an industry job as an escape
we cannot do academia well like
for example the google ai or facebook uh
ai research those are the they hire the
best
phds right uh so uh
like if you want to get into the right
very good
the promising industry research career
you need to be very good at
academia at research and creating
research that actually has an impact
so i guess those are my two points
yeah the ai is a unique area because the
company like
google amazon they tried to
hire the best ai scientists develop the
most advanced algorithms
but to develop the ai applications they
don't need a desire to have phds
yeah and i talk about that in my youtube
channel as well i feel like a lot of
industry application of machine learning
is actually an engineering problem not a
science problem you don't need to
develop algorithm to do that
you just need to apply the python
packages to self
you need you need to define the problem
do you need to apply the right python
packages to
to solve it
does anyone have any follow-up questions
yeah i would like to ask the final
question
okay yeah so the final one yeah
thank you for the talk it's a very uh
sensory
and content and it's your personal
experience
so my like first can you hear me
just to quick confirm okay so
my question is up
after that year after 6 years
phd everyone wants to
have that game more from the php
experience
so right now you choose to go to the
social industry and compete with a
master student and
some undergraduate students some of them
are your boss right now so how do you
how do you take this take this
situation how do you accept someone
is for it someone with only a master's
degree guide
to uh is your boss right now
i didn't uh kept the
kind of type maybe so i think he was
asking uh so right now when you enter in
the industry for example in facebook or
for example
your boss he only has a master's degree
how do you feel about that
yeah okay like
what's that meritocracy yeah like the
phd doesn't tell you you are better
right
having a phd header just uh is the
certification that you complete your phd
program
it doesn't mean you're better than
master so uh i guess if your boss is
better than you listen to him if
uh your bad idea is better than him you
try to convince him the idea is better
but i i don't view the title like the
title
the title does it shouldn't mean
anything
okay
yes i guess so we are coming to
one half hour and i really appreciate uh
eugene to take this penalty to share
the advice with oh is there any other
comment
uh
uh looks like there's another question
so this is definitely the last one so
uh i can paraphrase so he was asking
i don't know if the competition between
china and the us would inspire more
investments to fundamental
fundamental research uh uh what what is
your take on that
seems like right right now this is a
trend
i don't know and i don't care like what
can i do to change that
no so yeah
i agree on that uh so let's thank our
speaker again
for sharing uh his useful advice in
making these
decisions in throughout the phd and also
all uh
thank all the audience for coming today
and
especially for we are running uh this
talk series and that's talk is going to
be
uh centered around consulting so uh
if you say just uh follow our tweet uh
uh follow our uh wechat accounts we'll
be releasing announcements
and let's thank our speaker again and
hope you have a great weekend
all right thank you okay we can thank
you jim let's connect later
people who are interested in variety of
career tracks
and uh let's keep in touch
oh hey can can i also introduce myself
peace up here
sure go ahead oh sure yeah so uh so my
my name is uh and i'm now a phd student
in computer science in georgia tech and
yeah i i used to be a
master masters in university of michigan
and um
yeah i am glad to see you guys here and
um i'm never like i think i thought like
too seriously about what i'm going to do
in the future
but like so but like accidentally i got
added
in the weekend very glad and actually i
i saw some friends who like do
consulting jobs and
like it is a very appealing job for me
and i want to explore more
and see what it's like yeah thanks
it's good to see you i we have been we
shine friends for for a long time but
it's the first time i saw you
yeah yeah nice to see you
anyone else
you