|
Alexander L. Chobanyan |
|
Date of birth: February 17, 1977
Languages: Russian (Native), English
(fluent), Georgian (fluent), Armenian (understanding)
|
|
PROFILE ·
8 years of extensive statistical data
analysis and predictive modeling on large datasets ·
Expertise in building and tuning SAS high
performance forecasting process for business needs ·
Expertise in risk assessment models of long-range dependent random
time series ·
Expertise in statistical supervised learning methods for pattern
classification ·
In-depth knowledge of statistical techniques: ANOVA, Chi-Squared test,
parametric and nonparametric Regression, Principal Component Analysis, ARCH/GARCH,
feature selection and extraction, outlier detection, density estimation and
others ·
Solid programming experience in SAS (SAS Certified
Advanced Programmer for SAS 9.0), SQL, Matlab,
Java ·
Solid writing, presentation and communication skills |
|
|
STATISTICAL
FORECASTING RESEARCH EXPERIENCE / PROFESSIONAL EXPERIENCE
Business Forecasting Analyst, Logistics,
NESTLE USA, ·
Research analysis for adjusting SAS High Performance Forecasting
Process to NESTLE business needs ·
Development of SAS code modules to support statistical forecasting of
NESTLE USA sales Research Assistant, Department
of Computer Science and Engineering, Michigan State University Jan
2002 – May 2007 SUPERMEDIA Project: predictive modeling of tele-operation failure for supporting robotic control over
the Internet ·
Built
predictive random time-series models that describe behavior of available
bandwidth (spare capacity) of an end-to-end Internet path ·
Improved statistical kernel of present probe-rate based available bandwidth
measurement tools ·
Developed
self-verification algorithm that automatically adjusts modeling parameters
(such as a history window-size used for modeling/prediction and rate at which
network path “forgets” about its history) to a specific network
path ·
Investigated a
number of parametric heavy-tailed long-range dependent models used in
Insurance and Finance (ARCH/GARCH) and successfully applied one of them to
predictive modeling of an extremal behavior
(behavior of a minimum) of an available end-to-end bandwidth time-series ·
Applied
regression analysis and hypothesis testing to network measurements for
improvement of present probe-rate based available bandwidth measurement tools ·
Introduced non-parametric
approach to available bandwidth predictive modeling ·
Experimentally
verified applicability of time-series models and self-adjustment algorithm to
real-life datasets (in total, more than a million of real-life observations
from computers at Japan, China, Australia, Russia and US were collected and
analyzed) ·
Facilitated
team research activity between Electrical Engineering, Computer Science and
Probability & Statistics departments for carrying out SUPERMEDIA project ·
Presented at 5
conferences and published 4 papers Development of Anomaly Detection Systems for
Computer Networks ·
Analyzed
statistical patterns of wide-range of network denial-of-service attacks ·
Developed an initial set of 100 variables describing the state of a
computer network ·
Developed an
algorithm that chooses most relevant variables and creates network intrusion
risk model ·
Applied
statistical feature selection/extraction techniques (Log regression, ANOVA,
Principal Component Analysis and others) for figuring out final list of
network intrusion risk-sensitive variables ·
Applied number
of supervised learning pattern recognition statistical methods for describing
computer systems profile of normal operation together with
“alarm” thresholds ·
Developed Java
code for security-sensitive data extraction from network packets captured by
various network analyzers IT and Network Administration
Support, Department of Probability&Statistics,
MSU Jan
2001 – Dec 2001 ·
Developed and
maintained departmental website. Developed SQL queries for a web interface to
database ·
Planned,
implemented and maintained Microsoft Active Directory Infrastructure for a
computer network of the department ·
Was
responsible for internal security and virus protection issues of the
department IT and Network Administration
Support, United Nations (UNDP/UNV), Tbilisi, Republic
of Georgia Jan 1996
– Dec 2000 ·
Planned,
implemented and maintained Microsoft Windows NT based Infrastructure for a
computer network ·
Supervised all
computer/IT related budgeting, technical and management issues ·
Completed
“Leadership in Conflict Prevention and Management” course which
was organized by the National Peace Foundation, Software Developer, Georgian
Institute for Scientific and Technical Information, ·
Supervised a
group of five developers, guiding them with the creation of Georgian-English-Russian
CD “Business Georgia” EDUCATION
Ph.D. in Computer Science; Michigan
State University (MSU), ·
GPA: 4.0/4.0 Research projects: ·
Real-time
Data Streaming Across Wireless Noisy Channels. Used statistical analysis
of a recent packet-loss pattern to implement a number of modifications of Forward Error Correction protocol for wireless data transmission ·
Mobile
Agent based Intrusion Detection System (MAIDS). Implemented a prototype of Network Anomaly
Detection System based on mobile agents (Java Aglets) M.S. in Probability &
Statistics; Michigan State University, ·
GPA: 3.75/4.0 ·
Statistics
comprehensive examination: 126 out
of 128 M.S. in Physics (with honors); Tbilisi State University,
Tbilisi, Republic of Georgia July
1999 ·
GPA: 5.0/5.0 Certifications / Awards
·
SAS Certified
Advanced Programmer 2007 ·
Dissertation
Completion Fellowship; Michigan State University 2007 ·
Travel Grant;
Michigan State University 2006 ·
Microsoft
Certified Systems Engineer 2000 ·
Microsoft
Certified Professional + Internet 2000 ·
International
Prospective Student Scholarship sponsored by George Soros
Fund Management 1995-1999 ·
Honor
Scholarship of ·
Presentation
award for 2nd place at the International Soros
conference; ·
Presentation
award for 2nd place at the International Soros
conference; ·
Presentation
award for 2nd place at 55th Science conference for
Physics Students; ·
Award for
winning in personal rating (top 5%), Young Physicists Competition; List of selected
publications: ·
Behavior of Available End-to-end Bandwidth: Non-Parametric Approach ·
One-way Delay Trend Detection for Available
Bandwidth Measurement ·
Modeling Available Bandwidth for an
Efficient QoS Characterization of a Network Path ·
Intrusion Detection Systems Decision Logic Extension to
Detection of Sequential Anomalies |
|
Please see also attached Curriculum Vitae (Resume) Contact information: |
|