Personal
Statement
David
Chen
·
david
.
chen
@
berkeley
.
edu
· (415) 555-0192
PhD
in
Computer
Science
,
Fall
2027
The
first
time
I
understood
the
power
of
algorithms
was
not
in
a
classroom
,
but
in
a
cramped
back
office
in
Oakland
.
During
the
summer
of
2019,
I
watched
my
aunt
spend
three
evenings
manually
reconciling
inventory
logs
against
supplier
invoices
for
her
small
retail
business
.
I
wrote
a
Python
script
that
afternoon
to
automate
the
matching
process
using
approximate
string
matching
and
a
simple
hash
index
.
What
struck
me
was
not
merely
the
efficiency
gained
,
but
the
translation
of
a
tedious
human
process
into
a
formal
,
verifiable
computation
.
That
experience
taught
me
to
see
computer
science
not
as
a
vocational
skill
,
but
as
a
way
to
think
precisely
about
complex
systems
.
I
pursued
this
interest
at
the
University
of
California
,
Berkeley
,
where
I
completed
a
Bachelor
of
Science
in
Computer
Science
with
an
emphasis
in
systems
.
Coursework
in
Operating
Systems
,
Distributed
Systems
,
and
Algorithms
gave
me
a
rigorous
foundation
,
but
my
intellectual
home
became
the
research
lab
.
Under
the
guidance
of
Professor
Sarah
Reichelt
,
I
spent
two
years
studying
Byzantine
fault
tolerance
in
permissionless
networks
.
Our
project
,
which
we
called
BFT
-
Mesh
,
required
implementing
a
novel
gossip
protocol
in
Rust
that
reduced
message
complexity
from
quadratic
to
linearithmic
time
in
adversarial
settings
with
up
to
one
third
malicious
nodes
.
I
wrote
TLA
+
specifications
to
model
the
protocol
and
used
the
Ivy
verification
tool
to
prove
key
safety
properties
,
including
agreement
and
validity
under
partial
synchrony
.
This
work
was
published
at
the
2024
Workshop
on
Principles
of
Distributed
Systems
,
and
the
experience
taught
me
that
systems
research
demands
equal
parts
engineering
grit
and
mathematical
clarity
.
After
graduating
,
I
joined
Meridian
Data
as
a
software
engineer
on
the
consensus
team
.
There
,
I
worked
on
optimizing
Raft
log
replication
for
geo
-
distributed
databases
serving
financial
clients
.
The
engineering
challenges
were
immense
:
we
had
to
guarantee
linearizability
across
three
continents
while
keeping
tail
latency
under
ten
milliseconds
for
read
-
heavy
workloads
.
I
implemented
a
pipelined
batching
strategy
and
a
lease
-
based
read
optimization
that
improved
throughput
by
thirty
-
four
percent
in
production
benchmarks
and
reduced
cross
-
region
traffic
by
nearly
half
.
Yet
the
experience
also
exposed
a
gap
I
am
determined
to
close
.
Our
correctness
arguments
relied
heavily
on
manual
code
review
and
stress
testing
rather
than
formal
proof
.
When
edge
cases
emerged
in
production
,
they
were
often
subtle
violations
of
our
consistency
model
that
could
have
been
caught
earlier
with
better
verification
tools
.
This
tension
between
practical
scale
and
formal
guarantee
is
exactly
the
problem
I
want
to
spend
my
doctoral
research
addressing
.
1 / 2
A
PhD
in
Computer
Science
is
the
necessary
next
step
because
the
questions
I
care
about
cannot
be
answered
through
implementation
alone
.
I
need
deeper
training
in
randomized
algorithms
,
cryptography
,
and
program
verification
to
build
systems
that
are
provably
secure
under
precise
failure
models
.
I
am
particularly
drawn
to
research
at
the
intersection
of
distributed
systems
and
formal
methods
:
how
can
we
generate
verified
distributed
protocols
from
high
-
level
specifications
?
How
do
we
reason
about
the
composition
of
cryptographic
primitives
under
partial
synchrony
and
dynamic
membership
?
Can
we
create
automated
refinement
tools
that
connect
TLA
+
specifications
to
executable
code
with
machine
-
checked
proofs
?
These
questions
require
the
sustained
focus
,
theoretical
depth
,
and
collaborative
environment
that
only
a
doctoral
program
provides
.
I
am
especially
interested
in
programs
that
bridge
theory
and
systems
,
where
faculty
treat
verification
as
an
engineering
discipline
rather
than
a
purely
theoretical
exercise
.
I
want
to
work
in
a
community
that
values
building
real
artifacts
,
running
large
scale
experiments
,
and
subjecting
beautiful
ideas
to
the
harsh
constraints
of
production
deployment
.
My
industry
background
gives
me
a
respect
for
these
constraints
,
and
my
research
experience
gives
me
confidence
that
formal
methods
can
meet
them
if
we
ask
the
right
questions
.
My
long
-
term
goal
is
a
tenure
-
track
faculty
position
where
I
can
bridge
systems
implementation
with
formal
verification
.
I
want
to
build
tools
and
languages
that
make
verified
distributed
programming
accessible
to
working
engineers
,
lowering
the
barrier
between
research
advances
and
production
code
.
Beyond
the
lab
,
I
am
committed
to
mentoring
students
from
underrepresented
backgrounds
in
computer
science
,
continuing
work
I
began
as
a
teaching
assistant
for
CS
162
and
as
a
peer
adviser
at
the
Berkeley
Student
Learning
Center
.
The
problems
I
want
to
solve
require
patience
,
rigor
,
and
a
community
that
values
correctness
as
much
as
performance
.
I
am
eager
to
contribute
to
that
community
,
to
learn
from
faculty
and
peers
,
and
to
spend
the
next
several
years
working
on
systems
that
earn
trust
through
mathematical
clarity
rather
than
mere
testing
.
Thank
you
for
considering
my
application
.
2 / 2