CV
WORK EXPERIENCE
Intuit, Inc.
Staff Data Scientist (2021/2 - present)
Leading machine learning projects for identity and financial fraud prevention.
Mentoring junior team members.
Senior Data Scientist (2017/7 - 2021/1)
Developed new and enhanced machine learning models for real-time fraud detection.
Developed novel sampling techniques to avoid biases when training models on time-series data.
Contributed to the development of a central platform for real-time featurization with stateful aggregations on input data streams using Scala, Spark, and Redis, resulting in a patent filing.
Stanford University
Research Assitant (2011/9 - 2017/6)
Conducted academic research in the field of earthquake physics, using high-precision GPS data and physical models to study how the earth’s crust deforms before, during, and after earthquakes.
Analyzed GPS position time series from 1000 sites throughout Japan, measuring daily positions for 15 years (∼20 million data points). Modeled and removed short-term, spatially correlated trends and detected statistically significant long-term trends that were spatially coherent.
Analyzed space-time patterns of 1000s of small earthquakes in Japan. Applied statistical tests to detect trends in the recurrence times of sequences of collocated (“repeating”) earthquakes.
Developed tools for performing Bayesian inference and uncertainty quantification using Markov-chain Monte Carlo (MCMC) methods and applied them to a variety of linear and nonlinear inverse problems with different priors and/or bound constraints.
Splunk, Inc.
Data Science Intern (2016/5 - 2016/8)
Developed a graph-based model for detecting unknown malware in firewall logs (billions of events) using behavioral patterns in network traffic and label propagation from a smaller set of known malware.
Used Python for initial data exploration, then Scala and Spark for at-scale feature engineering, graph processing (using GraphX), and final end-to-end pipeline.
Achieved a detection rate that was 3 times higher compared to the baseline.
EDUCATION
Ph.D., Geophysics, Stanford University (2017)
Thesis title: "Long-term decrease in plate coupling prior to the 2011 Tohoku earthquake: Insights from GPS and seismicity data and physics-based models".
Advisor: Prof. Paul Segall
Cumulative GPA: 4.08/4.00
M.S., Geophysics, Stanford University (2014)
Cumulative GPA: 4.08/4.00
B.S. Magna Cum Laude, Geoscience, Cornell University (2011)
Honors thesis title: "Fault slip during the 2009 Northern Malawi earthquake sequence from InSAR data"
Thesis advisors: Matt Pritchard and Rowena Lohman
Graduating GPA: 3.85/4.00
SKILLS
Programming: Python, Scala, Spark, SQL, MATLAB, Java, R, Bash
Tools & Technologies: AWS (EMR, S3, SageMaker), Hive, Redis, Hadoop, HBase
Languages: English (fluent), Greek (native), French (beginner)
HONORS & AWARDS
Outstanding Student Paper Award (Seismology), AGU Fall Meeting, 2015
UNAVCO Geodetic Science Snapshot, 2015 [ link ]
AGU Eos Research Spotlight, 2014 [ link ]
Stanford Graduate Fellowship, 2012 – 2015
Cornell Engineering Research Honors, 2011
Engineering Learning Initiatives Undergraduate Research Award, Cornell University, 2011
Undergraduate Scholarship, Cyprus State Scholarship Foundation, 2007-2010
INVITED TALKS
Berkeley Seismological Laboratory Seminar, University of California, Berkeley, USA, September 1, 2015
US Geological Survey Earthquake Science Center Seminar, Menlo Park, California, USA, July 1, 2015
TEACHING EXPERIENCE
Stanford University
Teaching Assistant
GEOPHYS 288A: Crustal Deformation, Fall 2013 (Advanced graduate course)
GEOPHYS 90: Earthquakes and Volcanoes, Spring 2015 (Introductory undergraduate course)
ACADEMIC SERVICE
Reviewer: Earth and Planetary Science Letters (1), Journal of Geophysical Research: Solid Earth (2), Geophysical Journal International (1), International Association of Geodesy Symposia (1).
Student panel member, Geophysics Department faculty search, Stanford University, Spring 2016
PROFESSIONAL AFFILIATIONS
American Geophysical Union, Member, 2012―2017
PATENTS
Intuit IPR-2112053US, "Comparative Features for Machine Learning Based Classification", filed June 28, 2021 (co-inventor)
Intuit IPR-2010872US, ”Real Time Fault Tolerant Stateful Featurization”, filed May 29, 2020 (co-inventor)
PUBLICATIONS
See publications