My current research focuses on improving the effectiveness of adblocking and tracker-blocking technologies. Online advertisements have become an attractive target for various type of abuses, such as online tracking. Adblockers serve as a tool to protect user privacy by blocking these advertisements. Adblockers rely on filter lists to block advertisements. However, filter lists suffer from two major problems. First, they are manually curated with informal crowdsourced feedback and thus lack precision and accuracy. Second, manual curation adds an update overhead and make filter lists susceptible to evasion attacks. In my research, I address these challenges using machine learning approaches.
As an intern at Brave. I instrumented Chromium web browser to capture the rendering of a webpage. Specifically, I instrumented element creation, modification, removal, script execution, script compilation, and network requests initiation. The instrumentation provides sequence of rendering events for a webpage which can be generalized to a variety of problems.
As an intern at Microsoft, I wrote a technical report describing the current landscape of ad and tracker blockers. In light of developments around adblocking the report outlined how and what an adblocker for Microsoft Edge would look like.
As a solution analyst at LMKT Corporation., I worked on a number of projects. The most prominent projects were: (1) PTCL Smartlink: A mobile app for calling and instant messaging. It was packaged for PTCL (Pakistan Telecomunication Company Limited). (2) V-Govern: An e-governance solution. I added search functionality to the product with configurable similarity models. (3) RAFM (Revenue Assurance and Fraud management): A reporting dashboard. It provided near real time data analytics to monitor revenue and fraud critical situations by processing over one billion CDRs on daily basis.