Automation of Adverse Event Extraction
Center for Biologics Evaluation and Research (CBER) of
Food and Drug Administration (FDA) is tasked with ensuring the safety of biological products over the life cycle.
CBER medical officers review serious adverse events in the
Vaccine Adverse Event Reporting System (VAERS) within two business days. I have led a team that designed and implemented a clinical information extraction tool for
VAERS database.
Open Source Health Intelligence
The goal of this project is to build a cloud based system that uses open source data, such as social media feeds, to detect, monitor, and respond to outbreak events. I developed an information extraction module that was integrated into Splunk based dashboard of the system. The module extracts and summarizes information about morbidity, mortality, and hospitalizations, as well as geographic location of an outbreak event.
Document Categorization for OMIM
Online Mendelian Inheritance in ManŽ (OMIM) is a comprehensive database of genetic traits and disorders that is hosted and maintained collaboratively by the
National Center for Biotechnology Information (NCBI) and Johns Hopkins University. In this project I developed
an automatic method of assigning subtopic labels to
MEDLINEŽ documents relevant to a specific
OMIM record.
Mining and Analysis of Application Logs
Web analytics is of primary importance to
NCBI, since it develops and maintains more than hundred online applications which are accessed by over a million users every day. I analyzed application log data using pattern recognition and machine learning methods and provided recommendations
for improving usability of
NCBI applications.
Surveillance of Waterborne Outbreaks
This was
EPA funded public health surveillance pilot in the scope of
Water Security Initiative, which aimed to study the effect of incorporating
knowledge specific to waterborne diseases into syndromic surveillance system. Using semi-synthetic outbreak data I showed that adding water distribution schematic to surveillance system based on over-the-counter drug sales data improved the detection timeliness of waterborne outbreaks.
The goal of this
NSF funded
project was to develop a prototype of a mobile robotic assistant for elderly people at risk of institutionalization in order to allow them to maintain independent living for as long as possible. My part in the project was focused on passive monitoring of wellness. This was accomplished by sensing the physical parameters of activities such as gait and speech. I designed a biometric analysis system that processed speech and gait data and produced accurate data stream for the wellness monitoring module.
