►
From YouTube: How Bloomberg Uses AI to Help Speed up Over-the-Counter Trading - Camilo Ortiz & Philipp Meerkamp
Description
Don’t miss out! Join us at our upcoming event: KubeCon + CloudNativeCon Europe 2023 in Amsterdam, The Netherlands from April 17-21. Learn more at https://kubecon.io. The conference features presentations from developers and end users of Kubernetes, Prometheus, Envoy, and all of the other CNCF-hosted projects.
How Bloomberg Uses AI to Help Speed up Over-the-Counter Trading - Camilo Ortiz & Philipp Meerkamp
Did you know that trillions of dollars in securities are traded over-the-counter each year as a result of interactive conversations related to negotiated financial instrument transactions – also known as Dialogue Acts – conducted by traders using Bloomberg’s messaging tools (group chats and emails)? In this talk, we will describe how Bloomberg’s AI Engineering team builds and maintains services that detect offers of financial instruments in group chats to help traders who have requested to implement the add-on features that use these machine learning models to be more efficient in their negotiations. In particular, we will show how Kubernetes custom controllers – KServe and Kubeflow training operators – are critical in several steps of the model development life cycle (MDLC), including sampling of relevant data, training and distillation of models, A/B testing, and production deployments.
A
And
I'm,
presenting
with
Philip
how
we
use
kubernetes
in
particular,
to
speed
up
over-the-counter
Trading.
Hopefully
it's
a
little
bit
different
of
what
Ed
mentioned,
but
let's
see
how
it
goes.
Yeah
we've
been:
we've
been
in
Bloomberg
eight
years,
Philip
12
years,
mainly
working
on
NLP
and
AI
applications
in
Bloomberg.
A
So
the
idol
of
the
talk
I
will
try
to
do
like
a
crash.
Talk
on
crash
course
on
over-the-counter
trading.
What
does
that
mean?
How
Traders
or
people
in
the
finals
World
talk
to
each
other
and
Philip
will
go
over?
Why
be
able
to
detect
Security
offers
in
these
dialogues
is
important
and
then
we'll
try
to
bring
it
back
to
the
reason
of
this
conference
is
how
do
we
use
kubernetes
in
all
of
these
model
development
life
cycle
so
over
the
counter
trading?
A
In
particular,
if
you
look
at
the
stocks
or
the
normal
I
guess:
security,
some
products
that
are
traded
and
are
famous
I,
guess
among
the
people
that
are
not
into
Finance,
they
are
normally
traded
in
central
exchanges.
These
are
the
stocks
that
are
used
in
our
Pension
funds.
The
the
things
that
you
can
trade
in
your
apps,
but
normally
there's
like
a
central,
regulated
exchange
like
NASDAQ,
that
you
can
enable
that
people
can
basically
trade
and
they
will
see
like
a
single
price
for
everybody
to
to
check.
A
However,
and
I'll
give
more
stats
later.
A
vast
majority
of
the
financial
products
are
actually
traded
over
the
counter
or
traded
in
centralized
business
networks.
So
the
people
that
trade
effects
they
will
have
their
own
social
network
to
when
it
become
too
high
volumes
to
trade
for
an
exchange
same
thing
with
commodities
mortgages,
they
all
have
their
own
kind
of
social
network
where
they
talk
to
each
other.
A
Bloomberg
in
particular,
tries
to
bring
transparency
to
that
particular
market
and,
in
particular,
when
they
actually
talk
from
one
type
of
asset
class
to
another.
One
Bloomberg
tries
to
connect
all
of
those,
but
still
most
of
those
prices
are,
if
you
were
in
that
group
chat
or
in
that
social
network,
and
that's
basically,
what
over-the-counter
trading
means
is.
A
Basically,
when
you
have
when
people
trade
any
of
these
Financial
products
and
there's
no
supervision
of
an
exchange,
and
it
used
to
be
in
the
80s
that
you
will
see,
Traders
having
I,
don't
know
10
phone
lines
and
try
to
talk
with
everybody
and
repeat
it
exactly
the
same
offer
over
and
over
again,
you've
moved
to
emails,
but
it
quickly
became
a
Spam
problem
and
nowadays,
in
the
same
way
that
we
talk
to
our
family
and
friends.
Basically,
they
chat
with
each
other,
so
some
fun
facts.
A
Bonds.
Although
if
you
check
this
here,
it's
not
going
to
be
the
case,
but
bonds
has
historically
been
higher
than
Equity.
If
you
add
up
the
market
volume
and
most
of
those
bonds
are
actually
traded
over
the
counter
FX,
if
you
add
up
all
of
the
transactions
that
they,
the
the
Traders,
do
against
each
other.
They
are
normally
like
20
times
bigger
than
what
equities
traded
the
stocks
and
the
other
markets
and
they're
actually
also
in
the
high
volumes,
are
traded
over
the
counter.
A
And
if
you
just
count
the
number
of
things
that
can
be
traded
in
the
market.
Equity
only
accounts
for
1.5
percent
of
those
things
that
you
could
trade
at
different.
It's
only
like
30
000
at
most
equities
that
you
could
trade
and
ETFs,
but
all
the
other
products
like
municipal
bonds
and
structure
plugs
where
they
combine
I,
don't
know
like
a
bond
where
I
want
to
hatch
against
a
commodity.
All
of
those
are
actually
traded
by
people
talking
to
each
other.
A
A
In
addition
to
that,
in
the
first
post
we
could
detect
that
there's
an
inquiry
intent
and
there's
a
particular
bond,
which
you
need
to
link
that
we
have
a
link
in,
in
our
knowledge
basis
of
and
I
when
they
say
IBM
25s.
They
actually
normally
mean
at
that
point
in
time.
Ibm
coupon
is
seven
and
that
is
expires
in
October
30th
of
2029
2025.
A
and
the
size
when
they
say
10m
here
in
this
context,
is
actually
10
million.
They
want
10
million
of
those
notes.
If
we,
if
we
continue
the
second
post,
has
an
inquiry
intent
as
well,
but
a
different
size.
The
third
one
gives
a
quote
with
a
price.
Then
they
negotiate
more
negotiation
with
a
price.
A
trade
is
confirmed,
Bob
tries
to
negotiate
again,
but
then
there's
no
trade
right.
A
So
what
happens
really
behind
the
scenes,
so
it
could
have
been
that,
for
example,
in
this
whole
example,
where
that
Alice
checks,
the
news,
checks
the
portfolio
and
wants
to
get
get
some
of
those
bonds
when
Bob
or
Carl.
Whoever
serving
the
order
gets
that
inquiry
they
check.
What's
the
relationship
with
that
client,
they
check
all
of
their
RFQ
management,
they're
getting
quotes
for
or
increase
for
many
people,
and
they
also
see
what
will
be
the
price
that
they
can
give
and
they
give
a
price
same
thing
with
others.
A
They
jump
back
into
their
Bloomberg
or
to
their
Management
systems,
and
they
check
what
will
be
all
of
the
codes
that
they're
getting,
because
maybe
they're
not
only
talking
with
with
that
sales
person,
but
they
might
be
talking
with
other
people
and
they
will
just
decide.
Okay,
this
seems
to
be
a
good
price.
They
make
an
order
once
another
is
made
same
thing.
A
They
go
back
and
forward
from
from
one
system
to
the
unstructured
information
that
they
have
in
the
chat,
and
then
a
trade
is
done,
or
it
could
be
as
the
previous
example
that
they
take
too
long
and
and
it
passes,
so
what
we
really
wonder,
what
is
really
value
in
doing
dialogue?
Understanding
in
all
of
these
conversations
is
that
most
of
these
Traders
actually
have
200
active
group
chats
at
the
same
time.
A
A
B
B
We
can
recognize
here
that
apple
is
an
organization
and
Tim
Cook
is
a
person,
that's
the
recognition
part
and
further.
We
can
link
Apple
the
organization
to
Apple
Inc
in
California.
So
that's
a
database
entry,
an
identifier
in
a
bit
in
a
database
that
identifies
this
entity
apple
and
similarly
Tim
Cook
can
be
linked
to
Timothy
D
cook
in
Apple
Inc,
the
same
kind
of
named
entity
recognition
and
linking
you
can
also
perform
on
the
kinds
of
chats
that
Camila
will
just
talk
to
you
about
a
second
ago.
B
Similarly,
for
5.291
it
could
be
a
price,
it
could
be.
A
yield
turns
out,
it's
actually
a
coupon,
which
is
a
sort
of
an
interest
rate,
and
our
systems
would
be
able
to
identify
this
as
well
and
in
a
similar
fashion.
All
words
in
this
post
get
labeled
with
some
sort
of
type
and
all
together.
This
then
identifies
and
extracts
the
trade
that
is
being
discussed
in
this
IB
post.
B
B
A
So
I
will
just
glance
over
a
similar
cycle
that
we
have
talked
about
the
life
cycle
of
the
of
how
do
we
build
a
solution
like
this
in
Bloomberg
and
it
kind
of
blends
with
the
similar
workflows
that
we
have
seen
the
previous
talks,
but
it
all
starts
with
Gathering
data
right
and
to
just
illustrate
the
challenge
that
we
have
in
this
case
only
a
small
fractions
of
those
messages
or
those
posts
are
actually
mentioned
any
of
these
Securities.
So
a
lot
of
the
chatter
that
they
talk
to
each
other
is
building
relationships
right.
A
So
how
are
your
kids
doing?
How
should
we
have
coffee,
there's
a
lot
of
relationships
happening
and
every
now
and
then,
especially
when
the
Market
opens.
They
do
talk
about
this
this
Securities,
so
we
need
we're
forced
to
have
buyer
samples,
but
we
also
are
forced
to
track
how
what
was
that
biased,
The
annotation
task,
as
it
was
mentioned
before
by
Philip,
is
not
easy,
like
training
annotators
to
be
able
to
understand
those
dialogues
takes
us
a
long
time.
A
It's
a
big
effort
to
just
get
a
few
thousand
quality
annotations
for
for
a
given
asset
class
in
a
year
and
then
just
as
it
video
this
data.
So
most
of
these
posts
and
the
things
that
are
charted
in
this
in
this
in
this
in
these
groups
is
information
that
is
actually
hinting
of
how
the
market
is
going
to
move
five
minutes
from
there
or
one
hour
from
there.
A
So
this
we
need
to
make
sure
that
the
access
control
under
only
the
people
that
have
access
to
to
this
data
are
allowed
to
to
look
at
it
and
and
of
course,
it's
always
encrypted
and
at
rest
building
the
model.
So
even
if
we
have
perfect
data,
the
NLP
task
is
one
of
the
hardest.
It's
not
only
just
linking,
but
it's
also
a
slot
feeling
like
many
offers
can
be
mentioned
at
the
same
time.
A
A
Of
course,
since
we're
liable
on
a
lot
of
the
things
that
we
detect,
we
need
to
be
able
to
reproduce
and
explain
everything
that
we
detect
a
end-to-end
for
any
prediction
that
we
do
and
for
us
to
be
able
to
train
a
model
and
be
able
to
trust
that
evaluation.
We
need
to
account
for
the
bias
that
came
from
the
samples
and
for
most
of
the
knowledge
bases
that
we
linked
all
of
the
Securities.
A
Actually,
the
security
sometimes
only
spam
for
one
week,
they're
for
repurchase
agreements,
our
contracts
that
only
last
a
week
and
they
might
be
creative
within
the
day.
So
we
need
to
have
a
large
space
that
not
only
can
handle
that
in
real
time.
But
when
we
train
the
models
those
annotations
might
be
from
a
month
ago,
we
need
to
simulate
the
market
from
from
a
month
ago,
on
the
deployment,
the
usual
so
a
bunch
of
queries
per
second,
a
millisecond
latency,
depending
on
what's
the
the
application
and
how
much
automation
does
the
client
required?
A
A
So
I'm
not
going
to
go
into
the
Monitor
and
update
the
data
drift
detection
and
be
able
to
capture
the
feedback
and
be
able
to
create
new
data
based
on
that
feedback.
It's
another
set
of
problems,
but
the
theme
of
most
of
this
life
cycle
is
that
in
each
each
of
these
steps
we
need
to
have
all
of
the
dependencies
for
the
software
and
the
framework.
A
The
hardware
now
even
how
we
train
our
models,
change
depending
on
the
hardware
that
we
use
and,
of
course,
what
makes
it
even
more
difficult
in
ml
at
the
data
and
that's
where
we
basically
rely
in
our
machine
learning
platform
teams
to
using
crds
to
kind
of
normalize
how
we
do
this
across
on
this
model.
They
learn
life
cycle,
not
only
happens
for
security
detection,
but
when
we
want
to
do
intent,
classification
thread,
disentanglement
theme,
detection,
summarization,
you
name
it
right.
A
How
do
we
put
that
in
memory
deployment
out
of
all
of
the
crds
that
I
mentioned
K
service
is
the
only
one
that
is
open
sourced
where
we
deploy
our
models.
We
need
to
do
some
regression
tests
where
we
compare
a
lot
of
these
cases,
instances
against
the
previous
versions
and
monitor
an
update
where
we
want
to
do
some
data
drift
detection
and
a
stream
to
the
data
that
we
think
are
important
to
to
Gathering
feedback
on
overall,
and
this
is
solutions
that
I've
mentioned
with
kubeflow
and
Argo.
B
Historically,
the
mode
of
operation
in
machine
learning
has
been
that
you
would
gather
a
large
amount
of
annotations
thousands,
maybe
tens
of
thousands
which,
in
our
case
would
have
been
annotated
posts
where
a
human
painstakingly
annotates.
That
10
million,
for
example,
is
a
ticker
or
sorry.
It's
a
size
or
IBM
is
a
ticker.
B
B
B
B
So
overall,
this
lets
you
work
with
way
way
fewer
annotations.
Now,
how
does
that
connect
to
what
Camilo
just
said
for
one?
The
pipeline
that
we
have
at
the
bottom
of
the
screen
is
way
way
longer
than
the
pipeline
at
the
top
of
the
screen,
and
that
means
that
robustness,
reproducibility
and
the
kinds
of
things
that
Camillo
just
outlined
become
even
more
importantly,
it's
important
second
you're
working
with
large
models
and
again
kubernetes
and
the
ability
to
scale
work
on
in
multi-gpu
multi-machine
environments
effectively.
It
becomes
important.
B
So
Bloomberg
aims
to
provide
highly
accurate
data,
information,
analytics
and
insights
for
the
global
Capital
markets.
Ecosystem
human
interactions
still
dominate
a
significant
proportion
of
that
market.
The
over
the
counter
market
and
there's
an
opportunity
for
AI
to
make
relevant
connections
between
humans
who
operate
in
these
markets.
B
B
Our
team
is
growing
if
you're
interested
in
joining
us
do
scan
this
QR
code
and
you'll
be
taken
to
a
to
a
place
that
lists
our
open
positions
in
the
AI
and
ml
Ops
areas.
We
also
included
email
addresses
of
our
recruiters
here
for
both
London
and
New
York
City
positions,
and
we
have
a
Blog
Tech
AI
at
Tech,
bloomberg.com,
AI
and
Camilo
and
I
will
be
around
for
the
rest
of
the
day.
Don't
hesitate
to
talk
with
us
during
the
breaks,
we'll
be
excited
to
talk
with
you.
Thank
you.
C
Brilliant
yeah,
thank
you.
So
much
I'm
glad
you
did
the
talk
and
I
didn't
right.
It
was
my
summary
was
not
the
best,
but
what
an
interesting
problem
right
you
know
doing.
Nlp
on
multiple
chat.
Threads
seems
like
a
really
hard
thing
to
do
anyway,
but
where
the
stakes
are
that
people
are
trading,
tens
or
hundreds
of
millions
of
dollars,
you
really
do
have
to
get
it
right.
So
yeah
really
cool
use
case.
C
D
Yesterday
there
was
a
a
panel
or
sorry
an
entire
day,
focused
on
kubernetes
and
batch
and
I'm.
Just
it
didn't
really
go
much
in
the
details
into
your.
The
batch
I
was
wondering
if
you
could
maybe
talk
a
little
bit
about
like
how
you
guys,
if
you're
running,
like
batch
processing
on
kubernetes
and
what
you
have
found
with
it.
A
We
basically
have
our
own
crd
for
batch
processing,
so
in
a
way
each
of
the
teams
in
Bloomberg
they
have
their
own
namespace,
where
they
have
a
quota
so
that
they
don't
so
a
lot
of
resources
and
all
of
the
dependencies
are
put
in
a
yaml
as
usual,
and
then
the
the
batch
processes
is
run
for
python.
Tensorflow
spark
sometimes
so
most
of
these
projects
are
kind
of
managed
in
namespaces,
and
teams
share
the
results
of
the
experiments,
because
they're
all
running
the
same
in
the
same
share
workspace.