How to Survive in the Robotics Industry as a Software Developer

Gregory Pelcha

Recorded at GOTO 2016


Get notified about Gregory Pelcha

Sign up to a email when Gregory Pelcha publishes a new video

so this is a company that I work out
I've been there for a couple of years
now a little over three years and what I
want to talk about is the is the
industry itself and the mindset that it
tends to have I was one of the first
software engineers to actually join the
company but in order to kind of
understand the journey that I've been on
thought i would share a little bit about
the product so this is the founder and
CTO of our company and he was kind of
going through and he's setting up one of
our robots and the idea behind it is
just how simple it is to actually put it
together
so it's really just a few minutes of
actually going through and setting it up
it's basically a touchscreen with the PC
connected up to a robot and now he's
kind of going to go through and set up a
simple set of programs where it will run
across a two points in space running
he's showing off a little bit of the
user interface that we have here so you
can actually go through and you can
control the robot by using the the teach
pendant itself as which we call it which
is this tablet thing or he can push the
button on the back and then just start
moving the robot around it's not
remembering any of the motions that is
going through but once he stops it he
can store that information into the into
the program and then from here he can
actually start running the program
so he's kind of showing off a little bit
about the safety about it and some of
the many applications that we have
inside so you can do palletizing
stacking d stacking this is kind of
showing off some the precision of the
robot
it's actually a little bit funny to look
at this because this video is around
seven years old the robot itself is
roughly the same as it always has been
we've made some modifications to the
user interface but for the most part the
product is roughly the same but the big
selling points of the robot is how
friendly it is how flexible this is
something that you could actually pull
into your factory into an office setting
and because of the low noise and and
whatnot it's you can actually work next
to it in the video was talking about
about six to eight months worth of ROI
and now it's round five to six so the
way Universal robots views the robot
that we produce it's just the arm but we
see it as a tool as much as you would
use a hammer or a screwdriver this arm
is just a an extension of what you can
do with it of what a worker would do one
of the ideas around around the company's
belief system is is trying to solve some
of these these core problems so when you
deal with manual work you have variable
set of quality it's not going to be
consistent set of quality when when you
when you give it to your laborers and
the the overall cost of the wages and
what you can actually do with them tends
to vary so there's a possibility of
outsourcing so you can reduce some of
these costs but then what you're doing
is you're actually shipping out your
knowledge to another country and then
you run the risks of copyright
infringement sand and other issues so
automation is not a new thing it's been
around for a while the first generation
of automation was these great big
massive machines that you see in the
automotive industry they are big they're
very complicated to set up so therefore
they tend to be inflexible once you
place it to do a particular task they
just sort of stay there now you could go
through and you can get some specialized
machinery that would perform a
particular task but then the flexibility
is lost again
and then you also have to have this
specially made with 3d printing is
starting to become a little bit more
viable but it still doesn't solve all
the problems so we're a little bit
unique we're coming into a market that
has mainly been focused towards large
enterprise industries we're kind of
aimed at the small to medium and
therefore the market is quite rich and
if we start looking at the the industry
that itself there's actually been a huge
growth in the number of robots that are
being sold per year and a majority of
these are still the the large robots but
we're starting to see a change for
collaborative robots so these small
robots are starting to come into the
picture as well there was a small dip
here in 2009 but I go I guess we all
know what happened at that point that
was during the recession so we were
doing it ever we can to cut costs so if
you look at the different industries in
which robots are being used automotive
of course is one of the big ones and it
will probably be that way for for some
some time going forward we're starting
to see a lot in regards to electronics
especially in China and trying to get
automation around that and I think that
we're going to over the next 10 years
start to see other industries pop up
into this picture as well the one of the
challenges that you have in regards to
robotics and automation in general is is
the material that you work with if
you're working with clothing for example
it's so it's so fluid and and it's not
stiff so it's actually a challenge for
the the robot to do anything with it so
to put automation around that is quite
complicated but there was actually
someone in Seattle Washington that found
a way of stiffening the material and
making it so that you could use our
robot for sewing applications the
different types of applications that you
tend to see are really around the
handling welding and assembly again
there's still some growth that's going
on in this industry
and I think again we're going to start
seeing some some others pop into this if
we go back a little bit and start
looking at the the company itself when
it first started up it was really about
the product it was a couple of students
that had a dream of of getting into this
market and so they started working on a
prototype the you are five which is the
robot that sits in the middle is also
the one that was in the video and once
they kind of got that that prototype
into the to the picture they went ahead
and they started making their first
sales once that happened they shifted
over and it really started focusing on
the business and getting it out to the
entire world as as time went on we
started realizing that there was a need
for other robots within the family so
the you are 10 and the you are three
came about the number just represents
the number of kilos that it can lift the
you are three is a little bit special in
the fact that of all the robots it has
one joint that is capable of doing
infinite revolutions so we could
actually do screwing applications and
it's actually down at the tool so our
company it although it's only really
been in in the market around 2009 it's
actually been doubling every single year
and not only has it been doubling in
regards to the revenue has also been
doubling in the number of employees so
we've been seeing a huge number of
amount of growth which is actually kind
of exciting but we're starting to see a
little bit of a dip in that we're not
going to be doubling it's quite
difficult to maintain that that level of
that level of growth so we're all over
the world these are all the different
countries that we that we provide the
ones that are marked in blue and then
all the dots are the subdivisions that
we that we have our business model is
not that we actually sell directly to
the customers we actually have
distributors that we work with and
sell to the distributors and then the
distributors go and sell it out to the
end customers the nice thing is if we're
actually based here in Denmark we're
over Anna Lanza so that's where our
headquarters is and we have worldwide we
have roughly 300 people that are in the
company but here in the headquarters we
have about 200 people still continuing
to grow the special thing about it is
that everything in regards to the
company is done under one roof so we
have our R&D department there we do all
of our own development our own research
we have our own production they will go
through and do the assembly and
therefore we can have immediate feedback
in regards to the product that we make
we have the marketing and the sales and
everything there as well and we have a
tech support that provides corporate
side tech support which provides support
so this was a kind of an interesting
thing it was an honor for us this is
normally something that is offered to
celebrities but the Robo index was
opening up at that year and so we were
given the opportunity to ring the
closing bell I think it's kind of funny
the kid in the front he seems too
enthusiastic I'm not sure he knows
but the reasoning behind us being the
one there is that they wanted a robot
that was not behind doors in some
research and development facility they
wanted something that was actually out
in the industry so they chose our are
you or five now throughout the years
we've won a number of different awards
and we're very proud of those but this
one is kind of special not so much that
we're on this list from MIT but it's the
list of others that are on here as well
so you have the Tesla Motors you have
google you have Amazon you have SpaceX
and to actually be a part of these these
other companies that are doing some
pretty amazing things was it was a good
feeling good pat on the back for us we
have put a dash line around a competitor
there's sort of an interesting company I
think they're they're based in the US
and they have a dual armed robot can't
really say a whole lot about it it's
it's interesting it's a bit of a copy
from another robot that actually uses to
you or five arms but but they made it to
the list but we're still in the lead and
not too far back in time we were
purchased by a an American company and
I'm an American so that was actually
kind of yay great news but the the cool
thing about the the acquisition was that
they they really didn't want to come in
and make a lot of changes so they they
saw a company that was doing something
that that they can kind of get behind
and they just let us do what we do best
financial backing and moral support
so this is an actual application now I
said small to medium but businesses but
we focus on but BMW was so kind enough
to allow us to film what they do with
our robot and this was a a task that I
believe was being done by an individual
you can only imagine how boring that
would be but the robot doesn't care so
it's just going through and it's just
making this panel attached to the to the
door I'm not sure how many it does in
and now it'll switch over and show the
application again I think it's a
different door this is a different color
they've actually done a lot of
advertisement around this the fact that
they are using our robots in the
automotive industry and i think the the
other interesting thing about this is
that there's no fencing there's no
safeguards around it normally what you
would see in a plant is that you would
have the auto nation that's being done
with the robotics that's caged off and
people are not allowed to be in there
and then you have this other side where
the people are allowed to work and here
we're starting to see a blending of this
so the robots are beginning to work
side-by-side with the people this is not
a true application of that the people
they're kind of their own station the
robot is just simply out of station but
so my experience so I like I said
earlier I've been with the company for
about three years and when I joined
there was no software engineering there
was no processes in place in regards to
how we do development and it was mainly
these four categories of people that
were being hired and actually the
industry today is very much focused on
this you know in order to in order to be
in this industry you need to have some
robotics background you need to have
some mechanical engineering some
electrical engineering or you need to be
just really really smart with math and
coming into this company where it's
mainly master's degree and PhD students
that's the that's a little unnerving so
I was the only software engineer and
what I've been doing is trying to get
more software engineers into the picture
and i'll talk a little bit about more in
towards the end as to what the reasoning
is behind that but it's really to to
take on the software side of it
most of the software development that I
do is nothing really specific about the
robot it's about the user interface
about the user experience and when I was
hired there was no testers and there was
no UX designers it was just engineers of
what going through and making these
designs and so I wanted to change that I
wanted to put more of a I guess my spin
on it my my background so a typical a
typical way that they would work is that
you would be given a project from a
manager they would come up to you and
say this is your this is your feature
that you need to implement and I
remember my very first one it was
something called a flight recorder and
what it was responsible for is like the
last ten seconds of the robot run before
it crashed I needed to track all of this
information store it and make it
available for to be picked up later I
never implemented it but but that was
how it tended to work and they went to
the next engineer and they would hand
off another project next engineer and so
each individual had a project that they
were responsible for and to make sure
that everything was on track they would
have these meetings every two to three
weeks or every month depending on when
they felt it was necessary and we would
go through the entire rd department
whether it was someone that was doing
something on a computer board inside a
joint or whatever they would kind of
explain you know how their project is
doing and then it would come to me and I
would give my feedback as to how my
project is doing and that we would close
up and they will continue working for
the next month because it was handed off
to us this project we would do
self-testing so we would coat it up or
wire it up or whatever it is we did as a
member of the rd and then you would test
it that was our responsibility we will
also design it so if there was any sort
of a user interface or some sort of a
user interaction user experience we
would test that make sure that it was it
was good enough you could only imagine
what a room full of engineers could come
out
with the automation that was there was
kind of interesting so we had we had
automation around the the software so we
would commit our code it would go
through a build process we had a Hudson
build server that would go through and
compile it and produce some some
artifacts at the end of it we have this
thing called a Europe file and what that
really is for is to update the robot and
you could use a USB stick and plug it in
very old school and then you could
update the the robot to a newer version
the other thing that we have automation
around was kind of a new thing that I've
never seen in any other company was the
creation of the software manual and what
it would do is that we would have Java
running on top of the application and it
would go around from screen to screen
and it would take snapshots in multiple
languages and then that would be built
into a latex file system orale tech
build system and then a PDF would kick
out and from that PDF we would also
create some HTML and that would be
embedded into the Europe file there was
a bit of a chicken and an egg problem
there because we would build the
software before we would actually make
the manual that went with it and so
therefore you would have to kind of wait
until the next bill before he commits
so very shortly after I started I think
it was just a month month and a half I
sort of killed that idea of individuals
being responsible for a project and what
I did is I looked around at the the
software developers that were there
doing Java development and C++ and I
brought them into a small little team it
was a team of four maybe five but we
started doing development together we
would pool our features that we needed
to implement as individuals and we did
it as a team it was crumb that's
basically what I was implementing and
the company had never seen that before
but i think it was actually quite
helpful and instead of having these rd
meetings which we still continue to do
it was more of a team meeting on a daily
basis and of course we would do our
planning and her retrospectives and
demos was not really there yet I took a
little while longer but it did come
about but it was really about this team
dynamic very quickly I realized that we
need to test her badly and so I started
making some requests for this and the
company didn't really understand why we
would need this but we were getting some
students to come in and they said you
know why don't you take one of the
students and that will be your
designated tester okay it's only
part-time but we'll take it and that guy
was doing nothing but testing until he
until he graduated with his PhD and then
left right before he left he during one
of the stand-ups he just kind of made
this date me it's like I can't do it
anymore it's just too much work that
needs to be done so that's when we
started actually looking at hiring some
testers and now we have a couple of
the UX designer was kind of a funny one
my initial thought is that we shouldn't
be doing our own graphics I was actually
sitting there while I was coding and
trying to create icons and putting this
into the product also going through the
the web and trying to find free icon so
I can put him into our product as well
but what I saw is that we needed someone
to actually look at the user interface
and start improving upon it and the rd
director at the time said okay but they
also need to be a software developer so
they need to be capable of writing Java
but part time be able to do this ux
designer position and i started doing
the interviewing and just before I would
bring them in for an interview without
actually prep them like you need to
learn so a little bit of swing just go
online and just read up a little bits
and I'll do some quick questions with
you but don't worry I don't think you're
going to be doing anything in regards to
that management was very convinced that
would be doing at least like eighty
percent Java
development we hired our first person
and she never wrote a single line of
code and since then we've actually hired
a few more UX designers and what they've
been doing is is really trying to look
at the overall company and bring in all
of it so talking with marketing talking
with with sales tech support and using
that information that they have trying
to bring it back into Rd obviously we
started going through when creating a
some automation it was funny whenever we
wanted to deploy something out to a
robot we were dropping down to a command
line and then doing secure shell
commands from lionex pushing files over
to it one of the first things that I did
was actually create this simple script
it just pushed everything over but of
course there was also other things we
started doing some unit tests the very
first unit test brought up all of the
application and so we started decoupling
all of this and making it so that unit
tests were truly unit tests we started
introducing end-to-end test integration
tests and we started improving upon our
overall build process so that's one
aspect of how software engineers can
truly impact this industry but there's a
little bit more and if we go a little
bit into into the past what we saw was
that it was mainly the mechanical and
people the mechanical engineers that
that was really the the heart of of
automation there was no electronics no
digitization until just 40 years ago
during the the third Industrial
Revolution and then when that started to
happen we started to see software coming
into the picture and of course we throw
her hat into the ring because we were
kind of the first name regards to
collaborative robots there were other
companies that were trying to do it they
just couldn't do it in the way that they
were always doing the robotics industry
with these great big robots they wanted
to do this sort of like a waterfall
approach and they wanted to have all of
the safety and all the regulations in
place
that normal robots normally had em if we
go a little bit into the future what
we're beginning to see is that robots
are starting to invade into our daily
lives so it used to be that it was just
purely in the factory and now we're
going to start seeing it coming into a
lot of other places we've seen them in
the hospitals but I think we're going to
start seeing more into our homes so
we're starting to see the close of the
third Industrial Revolution that's
really about the digitizing of the
automation the first was steam and water
power the second was mass production the
third is the digitization but now today
we're starting to see cloud services
starting to come into the picture we're
going to start seeing the the blurring
of the lines between robots and and
humans and that endeavor is actually
kind of difficult if you're just
mechanic if you're a mechanical engineer
it's kind of difficult to to start
pushing things into the cloud and
starting to understand what does this
actually mean so that's where we start
to come into the picture right a really
cool one I think is this mass
customization so when you're going out
and you're you're getting a item of
product whatever you can have it
specially made just for you so if you
want it to be in a particular color with
gold stripes and and whatever you can do
that if you if you wanted to you know
have have other things on it is it's all
available to you what we're trying to do
is to embed information into the product
so as it's going through on the
production line it could be made in a
different way it could be made
customized for your needs now Universal
robots is very much geared towards
trying to
break down the walls of that's where you
do your automation this is where your
people are the do the to do the labor
and the main reason of that is that
you're kind of missing the human aspect
you can't you can't improve upon a
process that doesn't have people
involved in it and in order for to do
that you kind of need to bring humans
back into into the picture of being a
part of this automation to be working
with the robots so let them do what
they're actually good at which is the
creativity the thinking and then let the
robots do with their good at which is
the the automation and the quality and
and the repetitiveness so what we will
be kind of for see as the next
revolution in regards to the industry is
is really just empowering the the people
on the on the work floor so trying to
blur this line between blue-collar and
white-collar and actually empower these
people that are that are working on the
factory floor to to be able to make more
decisions that's kind of the idea behind
lean is so that they can react to the
things they can actually improve upon
and and bring some passion and love back
so if we look at the past in regards to
the into the industry with these great
big robots I mean there were pretty much
fixed installations now we would we want
to start looking at making them
relocatable so whenever you need to move
a robot to a new station we can do that
there's a number of companies out there
that will on monday and tuesday the
robot will be performing a particular
task but on thursday and friday they
need to move that robot over to another
station to do some finalization there's
a very interesting company in the US now
called hire buttocks and what they're
doing is actually outsourcing their
robots not out
sourcing but but replacing the concept
of temporary services so the idea is
that instead of hiring someone temporary
to do a manual task you hire a robot to
do a manual task probably talked a
little bit too quickly but really the
conclusion is is that I I foresee that
the software engineer is a very much
needed aspect within this industry and
I've been going through the the internet
and reading about some of the the
comments of like how do i get into
robotics and how do i be a part of this
and a lot of them are a lot of the
responses that i see are well you need
to have a robotics background and that's
not entirely true i don't have a
robotics background but i'm working for
our company and we're looking for more
people that are like me that have this
experience of developing software and
there was an article that was released
not too long ago that was talking about
the the fourth Industrial Revolution
this industry four-point-oh and there
were also mentioning the same thing you
need to have both you need to have an
understanding of the software side as
well as an understanding of the of the
robotic side and I don't think that's
entirely true a lot of the stuff that if
it is coming in the industry
four-point-oh is really about what
software engineers know how to do
they've been doing that for the past
almost decade now so we can we can take
our understandings our learnings from
that and actually bring that into the
industry thank you good Greg thanks
let's keep the guy hand
that's good we have van with time for
questions yeah probably the time that's
okay we can actually have a discussion
i'll start with the first question and
someone can follow up on that but when
you so when you got in as a software
engineer and you have you approaching
this team of hardware engineers yes
right is it you're thinking that
robotics hey these guys got to be wicked
cool and very fast iterations and agile
and the whole thing as you would think
of a software guy is that is that true
or is robotics engineers and firmware
developers like most hardware and
firmware developers are with other
products the I think it's a mindset
thing but yeah for the most part it is
is not having the the really quick
iterations one of the one of the
challenges you have when you're dealing
with hardware of course is that you you
come up with a design and then you have
to wait for that design to be
materialized now 3d printing is a
possibility that can give you some fast
feedback but it's not the same as
software you can't just release that out
into the world and get it and get the
feedback the other thing is that the the
way customers tend to to look at
robotics is a little bit different as
well they love stability and therefore
they don't want any change we have some
robots that are sitting over it at Lego
they're not doing any sort of production
work they're just doing testing but
they're still running on an extremely
old set of software they had no need to
update the software it just constantly
does the same thing over and over again
and I think when it comes to most of the
the thinking around the customers out
there if they don't they don't foresee a
need and we as much as we may encourage
them to to want to do these updates if
it's work it's if it's working you know
just just let it be so it's a
it's difficult yeah and any other any
questions here yeah told him thank you
Sarah nice wonderful group so I'll do
the same because Greg night know each
other's well so so what so you had this
slide where the way it was before and
then you change stuff and yeah it's a
room full of software developers and we
all look at and say well you turned a
bad process designs probably a better
process but I don't know if it was a I
don't the tangible results that your
customers feel a difference that's
actually a very good question that's
difficult to answer it's it's not
something that we actually have a lot of
feedback on one of the problems that I
think that we have with our business
model although the business model is
very successful in regards to the
business side is that the feedback loop
is actually quite long so a customer
will report to subdivision and in the
subdivision might report to to tech
support and then tech support will
report into Rd and actually there's a
little bit of a that's not entirely true
because the end customer will most
likely go to a distributor and try and
get support then it goes to the
subdivision then to tech support then to
us so you have this very long feedback
loop so it's difficult for us to to be
able to to get any real measurements as
to do they feel like an improvement of
the quality or are they happy with but
that we have dedicated team that is
dealing with supporting external
customers they're a little bit happier
so the way it worked before it kind of
went into the R&D department and then it
might sit there for a while and then we
would provide some some update I think I
think one of the real benefits that we
get is the is the synergy of the team
and it's not just one individual that's
trying to dream up you know what does
this look like and how do I test it and
and what's the user experience going to
be like instead we have a team that's
actually rallying around us and so from
that perspective I think there's been an
improvement so but it is difficult to
measure good any other questions I have
one okay so we like to think that
software is eating the world especially
in sort of old or industries I think
that's fair to say that's somewhat true
is that the case in the robotics
industry as well as software changing
the way the industry is B is delivering
value and if so how well that's a tough
one I think it is changing it I'm not
really sure how to even answer that
question you have to realize that I came
into the into the robotics industry and
all I saw was software I think it's
interesting because robotics is sort of
born into the digital age right defining
the fourth generation of the Industrial
Revolution so is software and not sort
of taken for granted you would think
that it's embodied inside the industry
yeah so to me it just it sounds weird
that you that I mean that you have to
survive inside a robotics industry yes I
love the title I didn't come up with the
title I i proposed to them the what my
talk was going to be about and it's not
really survive but regardless like amy
was the first guy right you were the
first time i was a searched one and I I
think he had some big plans as to some
software challenges specifically that he
wanted me to take on and those have not
come about but a lot of the time and
energy has been spent around trying to
improve the overall development process
eating the world I've never heard that
term well you just heard it yeah yeah
trying to internalize that a little bit
here but yeah I i I'm not really sure
that's difficult to say all right any
other questions say if not we will give
you an early leave thanks a lot for