Calibrations on the Web for one of the World's largest Engine Manufacturers
# Improved engine calibration download and usage performance for the client from a few hours to less than thirty seconds using Hadoop as an Infrastructure-as-a-service model on the cloud.
# Designed and documented the end-to-end solution for the “Calibrations on the Web” project which put calibrations on a cloud and extracted them at the user’s wish.
# Lead a team of developers who implemented the entire solution for “Calibrations on the Web” and guided them to improve BigSQL performance from about a minute to less than thirty seconds.
FriendVoice: Socially Smart Product Reviews and Analytics
# Developed a socially smart product reviews and analytics system which integrates social data from Facebook to enhance user experience by filtering reviews according to social relevance.
# Developed algorithm for helping businesses to use this data to gain insights to their customer base and direct specific marketing efforts to individuals and geographies.
# Did detailed market research on existing mobile applications which help users to plan trips with public transport systems.
# Based on it came up with a solution for designing a Hybrid mobile application using Phone Gap which could help the local Pittsburgh people to plan their trips.
# Also did estimates on the cost of development, deployment and maintenance. And proposed ways of generating revenue from the application.
# Designed end to end software model for a mobile banking application, from requirements gathering to detailed architecture design; and presented to management for funding approval.
# Developed a detailed spec which included UML diagrams such as ER diagram, Sequence diagram and Class Diagram.
# Conducted contextual inquiry to gather usability perspective for existing drug stores like RiteAid and Walgreens and performed its usability study.
# Based on the study, designed an eCommerce web portal mock-up and performed its heuristic analysis.
Fraud Detection for one of Canada's "Big Five" Banks
# Improved the fraud detection capabilities of the customer, reducing the false positives by about 90% of the original statistics.
# Used IBM BigSQL to analyze Hive tables to detect fraud transactions in customer data and used the HBase Java API to join data from multiple HBase tables; and did performance analysis of each case.
Lecturer for Java and Computer Networks
Taught the Java programming Language and Computer Networks to Undergrad Computer Science Engineering students
Secure eVoting System
Developed a system for securing voting mechanism by making use of Secure Multi-party Computation and Virtual Party Protocol. Voters login and vote for their chosen party and data is sent to virtual parties for security and then to Trusted Third Party for final calculation.
InfoSphere BigInsights guidance and advisory to one of Canada's "Big Five" banks
# Advised the customer on possible solutions to off-load data from EDW systems into InfoSphere BigInsights; and designed and documented the end-to-end solution.
# Guided the customer on Operations Procedures in InfoSphere BigInsights for administrative tasks.
# Gave demo of several components of InfoSphere BigInsights to the customer including BigSQL security, Data Federation, Data loading and Ingestion, Oozie, Monitoring in BigInsights, and NFS mounts for GPFS.