Research Experience

Role: Graduate Research Assistant
Location: PurSec , Purdue University
Duration: Spring 2022 - Current

📌 Project: Investigating user perception of UI attacks in WebXR within the advertising ecosystem.

🔹 Identified five novel UI attacks within WebXR ad ecosystem. Proposed a four-category taxonomy for 14 such attacks based on objectives of the adversaries.

• We leverage dark pattern attributes (e.g., Covert, Deceptive, Restrictive, Information Hiding) to identify 14 security sensitive UI properties that contribute to these UI attacks.

🔹 Developed a 3D spatial log framework and four quantitative interaction metrics to assess user engagement within WebXR environments.

• The log framework captures various events such as human intended or human unintended/uncontrolled click or focus on task and dark pattern objects, position of intersection, cursor movement information, camera position, direction etc.
• Employing the log data, we measure presence, safe engagement, malicious attention and blind spot rendering fraction metrics.

🔹 Conducted a 100-participant in-lab between-subjects user study to assess user perceptions of the four attack categories within our taxonomy.

• We apply qualitative and quantitative analysis to assess the impact of UI attack categories on user engagement and active participation in a given task.
• We find that these attack categories are effective in achieving their malicious objectives while most of these categories go largely unnoticed by users.

🔹 Available Research Artifacts
🔹 Tech stack - A-Frame, Three.js, Javascript, Python, R. ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

📌 Project: Secure group pairing of co-located Mixed Reality (MR) headsets addressing potential adversarial threats.

🔹 Developed a novel localization system for pairing MR headsets using eye-tracking, hand-tracking sensor signals and spatial anchors.

🔹 Designed a high-entropy random hand gesture generator by anchoring a 2D gesture grid in world coordinates and detecting hand positions from the camera view.

🔹 Designed a CNN-LSTM network leveraging eye-tracking and IMU sensor data to detect synthetic data and secure pairing against adaptive adversaries.

🔹 Designed and conducted in-lab user studies to evaluate system success rate, scalability, and usability.

🔹 Tech stack - Standalone Mixed Reality app using C#, MRTK, and Unity, Python. ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

📌 Project: GPU based side-channel attack in XR.

🔹 Identified low-resolution GPU metrics related to object rendering in XR.

🔹 Fingerprinted 100 WebXR and 100 standalone XR apps, along with various XR objects within the apps, achieving over 90% accuracy using classical ML and DL models (e.g., Random Forest, SVM, CNN, LSTM).

🔹 Tech stack - XR app development using C# and Unity, Python.

Role: Undergraduate Research Engineer
Location: MCNRG, NIT Durgapur
Duration: April 2018 - June 2019
ITRA Project, in collaboration with Kalyani Government Engineering College.

📌 Project: Post Disaster Situation Analysis and Resource Management Using Delay-Tolerant Peer-to-Peer Wireless Networks.

Goal: To visualize the summary of GIS data shared by victims and volunteers in a disaster-stricken area in offline mode.

🔹 Developed a system for continuous parsing of user-obtained GIS data (via offline syncing, not internet) and storing it in a database. Built the “Surakshit Dashboard” to provide periodic summaries of collected data (polygon, text, image, video). The summarized data assists the rescue team by providing crucial information to aid victims after a natural disaster.

🔹 GitHub Link
🔹 Tech Stack - JAVA, Python, JavaScript, MongoDB

Role: Research Intern
Duration: May 2018 - June 2018
Location: CNeRG, IIT KGP
CNeRG Internship

📌 Project: Peer-to-Peer Live Video Streaming based on Scalable Video Coding.

Goal: The primary objective was to optimize video quality on the client side based on network conditions and display resolution capabilities while minimizing server load.

🔹 Developed a system to encode .mp4 files to .svc, transfer them using BitTorrent-like communication between server and client, and decode them back to .mp4.

🔹 GitHub Link
🔹 Tech Stack - JAVA, Python