Data Computer Science

JAYANT GUPCHUP

*******@***.*** 240-***-**** http://www.cs.jhu.edu/~gupchup

EDUCATION

Johns Hopkins University Baltimore, MD

Ph.D. in Computer Science, Candidate Expected: Jan 2012

Dissertation: Data Management of Environmental Monitoring Sensor Networks

Advisors: Prof. Alex Szalay and Prof. Andreas Terzis

M.S. in Applied Mathematics and Statistics, GPA: 3.67 May 2011

Thesis: Classification-Based Event Detection in Environmental Networks

M.S. in Computer Science, GPA: 3.75 Dec 2007

Mumbai University, Veermata Jijabai Technological Institute (VJTI) Mumbai, India

B.E. in Computer Engineering, Top 10% of class Jun 2003

Thesis: Data Mining and Indexing of Astronomy Databases

RELEVANT COURSEWORK

Computer Science:

Algorithms, Parallel Programming, Transaction Systems,

Information Retrieval, Natural Language Processing

Applied Statistics and Mathematics:

Data Mining, Statistics, Time Series Analysis,

Linear Optimization, Matrix Analysis

SKILLS

Programming: Java, C#, C, C++, Perl, Transact-SQL, JavaScript

Software: MATLAB, Excel/VBA, ArcGIS

Environments: Hadoop, Amazon Elastic Map Reduce

Data Analysis: Robust and Streaming Principal Component Analysis (PCA), Robust Regression (LLSE),

Classifiers for Categorical & Numeric Data, Expectation Maximization (EM), Hierarchical

Clustering, Hidden Markov Models (HMM)

PROFESSIONAL EXPERIENCE

Johns Hopkins University Baltimore, MD

Research Assistant, eScience & Sensor Networks Group Aug 2007 - present

Timestamping algorithms: Developed two energy-efficient algorithms to assign timestamps to sensor

measurements. Developed in Java.

- Resulted in 10% increase in data yield and enabled networks to operate unattended for months.

- Implemented a sensor network simulator to understand scalability of the solution.

Environmental data management: Designed and implemented a multi-stage end-to-end data processing

pipeline for sensor data streams in Transact-SQL and C#

- Current Size: 100 Million+ records.

Analytics of network logs: Analyzed and improved the performance of a wireless sensor network based on

network logs

- Identified failures causes, metrics and performed statistical analysis to predict failures of components.

Sensor data visualization: Designed a system to visualize real-time, multi-resolution environmental data

collected by the JHU soil sensor network.

- Used by the LifeUnderYourFeet soil-ecology group at the Johns Hopkins University.

- Mentored two masters and two undergraduate students.

Data-driven data collection: Improving the data collection system by exploring the tradeoff between

communication costs and fidelity using robust statistical methods.

- Developed an incremental and robust principal component analysis (PCA) library in MATLAB.

PROFESSIONAL EXPERIENCE (CONTINUED)

Microsoft Research San Francisco, CA

Research Intern, Scalable Servers & eScience Group Jun 2007 Aug 2007

Matlab access to DataCube: Implemented a data adapter and MDX query builder in MATLAB to access data

from a remote SQL Analysis Server.

- Used by the AmeriFlux scientific collaboration.

Inter-University Centre for Astronomy and Astrophysics Pune, India

Project Officer, Virtual Observatory Project Jun 2003 Aug 2005

Telescope data archiving: Part of a team of five that designed an archival system for data generated from

optical telescopes. Developed in Java

Data exploration of astronomy surveys: Employed SQL query plans to retrieve data for analytics of the Sloan

Digital Sky Survey database (~ Tb)

SELECTED PUBLICATIONS AND TALKS

J. Gupchup, D. Carlson, R. Musaloiu-E, A. Szalay, A. Terzis. Phoenix: An Epidemic Approach to Time

Reconstruction, 7th European Conference on Wireless Sensor Networks, 2010, Portugal. Oral presentation.

J. Gupchup, R. Musaloiu-E, A. Szalay, A. Terzis. Sundial: Using Sunlight to Reconstruct Global Timestamps,

6th European Conference on Wireless Sensor Networks, 2009, Ireland. Oral presentation.

J. Gupchup, R. Mus loiu-E, M. Chang, A. Szalay, K. Szlavecz, A. Terzis. Deploying Advanced Wireless Sensor

Networks for Ecological Monitoring, Microsoft eScience Workshop 2008. Oral presentation.

J. Gupchup, C. V. Ingen. Enabling Eco-Science Analysis with Matlab and DataCubes in the Cloud, Microsoft

Technical Report, 2008.

VOLUNTEER EXPERIENCE

Harvard Humanitarian Initiative

Developer, Haiti Voices Jan 2010 July 2010

Data management portal for Haiti relief: Developed a data entry and search portal using Google

visualization API to provide real-time information to non-profit organizations working in Haiti.

INTERESTS

Data visualization, Guitar playing

REFERENCES

Prof. Carey Priebe

Prof. Alex Szalay

Applied Mathematics & Statistics

Department of Physics & Computer Science

Johns Hopkins University

Johns Hopkins University

Phone: 410-***-****

Phone: 410-***-****

Email: ******@***.***

Email: ******@***.***

Dr. Razvan Musaloiu-E. Dr. Catharine Van Ingen

Software Engineer, Partner Architect

Google Microsoft Research

Phone: 650-***-**** Phone: 425-***-****

Email: *******@*****.*** Email: ********@*********.***

Contact this candidate