Research

• An Intelligent and Unified Log Analytics Infrastructure (Ongoing) 

  Current ongoing R&D project. The system aims to get insights from application, system logs and user generated events inside a DMZ to facilitate sensitive data leakage prevention (PHI, HIPAA), understanding usage trends, resource provisioning, problem diagnosis etc.

• An Automatic Method for Red-eye Detection and Correction in Digital Images (Undergrad Thesis) 

  Developed a state of the art automatic red-eye detection and correction system based on image processing and pattern recognition algorithms. The system achieved above 90% accuracy rate with very fast response as the system does not rely on Machine Learning.

• Web Application Security and Code Quality Analysis with Static Code Analyzers 

  Used multiple open source and proprietary static code analysis tools (Sonarqube, Checkstyle, PMD, FindBugs/SpotBugs) on the codebase of multiple platforms (Web, Android, iOS) to detect potential vulnerabilities (OWASP Top 10 Security Vulnerabilities). Later fixed those findings using Pareto Principle (80% of consequences come from 20% of the causes) which fixed more than 95% of the reported issues.

• Sandman - An Intelligent Sleep Tracker 

  Developed a sleep quality prediction system based on the features extracted from Heart-rate, Movement, and Circadian Rhythm data from the MEMS and PPG sensors of Apple Watch. Sleep stage detection was done using Random Forests and Neural Networks (Multilayer Perceptron) which achieved 90% accuracy and 60% specificity.

• REST API Request Profiling and Performance Analysis in Highly Concurrent Environments 

  Conducted load testing with Gatling on REST API endpoints to profile API performances from various metrics such as throughput, response time, Database query performance, concurrency handling etc. Later optimized the APIs to achieve accepted level of performance.

• Performance Comparison of AODV and DSDV Routing Protocols in Mobile Ad Hoc Networks (MANET) using NS2 

  There are various routing protocols available for MANETs. The most popular ones are DSR, AODV and DSDV. In this project, I simulated two routing protocols for mobile ad hoc networks– the Destination Sequenced Distance Vector (DSDV) and the Ad hoc On-Demand Distance Vector routing (AODV) and evaluated both protocols based on packet delivery fraction and throughput while varying number of sources and pause time.