Priya
Srinivasan
Machine
Learning
Engineer
📧
priya
.
srinivasan
@
email
.
com
📞
+1 (415) 832-9174
📍
San
Francisco
,
CA
🔗
linkedin
.
com
/
in
/
priyasrinivasan
🐙
github
.
com
/
priyasrinivasan
Machine
Learning
Engineer
with
5+
years
of
experience
building
production
ML
systems
.
Specialize
in
NLP
,
recommendation
systems
,
and
MLOps
pipelines
.
Passionate
about
translating
research
into
scalable
,
maintainable
machine
learning
products
that
drive
measurable
business
impact
.
EXPERIENCE
Senior
Machine
Learning
Engineer
Lumina
AI
Aug
2022 –
Present
Machine
Learning
Engineer
Vantage
Analytics
Jan
2020 –
Jul
2022
Data
Scientist
ClearPath
Health
Jun
2018 –
Dec
2019
EDUCATION
M
.
S
.
in
Computer
Science
Stanford
University
Sep
2016 –
Jun
2018
Specialization
in
Machine
Learning
&
AI
.
GPA
: 3.92.
Thesis
:
Efficient
Attention
Mechanisms
for
Long
-
Document
Understanding
.
B
.
Tech
in
Computer
Science
Indian
Institute
of
Technology
,
Madras
Jul
2012 –
May
2016
GPA
: 9.1/10.
Recipient
of
Institute
Silver
Medal
.
Minors
in
Mathematics
.
SKILLS
ML
&
Deep
Learning
MLOps
&
Infrastructure
Designed
and
deployed
a
real
-
time
content
recommendation
system
serving
4
M
+
daily
active
users
,
increasing
engagement
by
23%
and
session
duration
by
18%
·
Architected
a
multi
-
stage
ML
pipeline
using
TensorFlow
Extended
and
Apache
Beam
,
reducing
model
training
time
from
8
hours
to
45
minutes
·
Led
migration
of
batch
inference
pipeline
to
online
serving
with
TensorFlow
Serving
,
achieving
p
99
latency
under
35
ms
with
1200
requests
per
second
·
Built
an
automated
A
/
B
testing
framework
for
model
evaluation
,
reducing
experiment
cycle
time
from
3
weeks
to
4
days
·
Mentored
3
junior
engineers
through
code
reviews
,
design
documents
,
and
weekly
pairing
sessions
·
Developed
a
named
entity
recognition
pipeline
for
legal
documents
using
a
fine
-
tuned
BERT
model
,
improving
extraction
accuracy
from
82%
to
94%
·
Built
a
fraud
detection
ensemble
(
XGBoost
+
neural
network
)
processing
2.5
M
transactions
daily
,
reducing
false
positive
rate
by
34%
while
maintaining
recall
above
96%
·
Implemented
feature
store
using
Feast
,
centralizing
200+
features
across
6
teams
and
reducing
feature
engineering
duplication
by
60%
·
Created
model
monitoring
dashboards
with
Prometheus
and
Grafana
,
enabling
real
-
time
drift
detection
and
automated
alerting
·
Built
a
patient
readmission
risk
prediction
model
using
gradient
-
boosted
trees
and
claims
data
,
achieving
AUC
of
0.89
and
reducing
30-
day
readmissions
by
12%
·
Created
an
NLP
system
to
extract
medication
timelines
from
unstructured
clinical
notes
,
achieving
91%
F
1
on
a
held
-
out
test
set
·
Developed
ETL
pipelines
in
Python
and
SQL
to
aggregate
data
from
3
disparate
EHR
systems
into
a
unified
analytics
schema
·
1 / 2
TensorFlow
,
PyTorch
,
XGBoost
,
scikit
-
learn
,
Transformers
,
Hugging
Face
,
ONNX
,
LightGBM
Programming
Python
(
expert
),
SQL
(
advanced
),
C
++ (
intermediate
),
Java
(
intermediate
),
Go
(
basic
)
Kubernetes
,
Docker
,
TFX
,
MLflow
,
Kubeflow
,
Feast
,
Airflow
,
Prometheus
,
Grafana
Data
&
Cloud
Apache
Spark
,
Apache
Beam
,
BigQuery
,
Snowflake
,
AWS
(
SageMaker
,
ECR
,
EKS
),
GCP
(
Vertex
AI
,
Dataflow
)
PUBLICATIONS
&
TALKS
PROJECTS
LLM
Fine
-
Tuning
Toolkit
•
PyTorch
,
LoRA
,
TRL
,
Weights
&
Biases
Open
-
source
library
for
parameter
-
efficient
fine
-
tuning
of
large
language
models
,
supporting
LoRA
,
QLoRA
,
and
AdaLoRA
. 1,200+
GitHub
stars
.
Used
by
3
research
labs
for
production
deployments
.
DriftGuard
•
Python
,
Evidently
AI
,
FastAPI
,
PostgreSQL
Real
-
time
model
monitoring
service
that
detects
data
drift
,
concept
drift
,
and
performance
degradation
.
Includes
automated
retriggering
of
training
pipelines
and
Slack
alerting
.
Deployed
across
4
production
models
.
GraphRec
Simulator
•
PyTorch
Geometric
,
DGL
,
NetworkX
Research
framework
for
experimenting
with
graph
neural
network
-
based
recommendation
algorithms
.
Reproduced
state
-
of
-
the
-
art
results
on
3
benchmark
datasets
and
contributed
2
novel
aggregation
methods
.
CERTIFICATIONS
AWS
Certified
Machine
Learning
—
Specialty
(2023)
Google
Professional
Machine
Learning
Engineer
(2022)
TensorFlow
Developer
Certificate
(2021)
Srinivasan
,
P
.,
Chen
,
L
., &
Kim
,
J
. "
Efficient
Long
-
Document
Understanding
via
Sparse
Attention
."
NeurIPS
2023.
·
Srinivasan
,
P
. "
Building
Production
-
Ready
Recommendation
Systems
."
Talk
at
MLConf
San
Francisco
2024.
·
Patel
,
R
.,
Srinivasan
,
P
.,
et
al
. "
Real
-
Time
Fraud
Detection
at
Scale
."
KDD
Applied
Data
Science
Track
2022.
·
2 / 2
