I am currently pursuing a Master's in Computer Science and Engineering from Texas A & M University, College Station with a
specialization in Artificial Interlligence (AI) and Machine Learning (ML).
My coursework at Texas A&M revolves around the following specializations:
Location: College Station, Texas
Duration: Aug 2021 - Present
Program: Master of Science (Thesis) - Computer Science
Location: Chhattisgarh, India
Duration: Aug 2016 - July 2020
Program: Bachelor of Technology - Computer Science
CGPA: 9.30 / 10.00 (10.00 is best)
Location: Singapore
Duration: Jan 2020 - Aug 2020
Role: Research Intern (Machine Learning)
Location: London
Duration: Mar 2021 - July 2021
Role: Consulting Research Engineer (Remote)
Location: Bengaluru, India
Duration: Aug 2020 - July 2021
Role: Business Analyst
Abstract: The recently proposed DNS-over-HTTPS (DoH) protocol is becoming increasingly popular in addressing the privacy concerns of exchanging plain-text DNS messages over potentially malicious transit networks (e.g., mass surveillance at ISPs). By employing HTTPS to encrypt DNS communications, DoH traffic inherently becomes indistinguishable from regular encrypted Web traffic, rendering active disruption (e.g., downgrading to the plain-text DNS) by transit networks extremely hard. In this work, we investigate whether DoH traffic is indeed indistinguishable from encrypted Web traffic. To this end, we collect several DoH traffic traces corresponding to 25 resolvers (including major ones, e.g., Google and Cloudflare) by visiting thousands of domains in Alexa’s list of top-ranked websites at different geographical locations and environments. Based on the collected traffic, we train a machine learning model to classify HTTPS traffic as either Web or DoH. With our DoH identification model in place, we show that an authoritarian ISP can identify ∼97.4% (∼90%) of the DoH packets correctly in a closed-world (open-world) setting while only misclassifying 1 in 10,000 Web packets. To counter this DoH identification model, we propose an effective mitigation technique, making the identification model impractical for ISPs to filter and consequently downgrade DoH to plain-text DNS communications.
Abstract: In recent years, the advancements in Information and Communication Technology (ICT) are mainly focused on the Internet of Things (IoT). In a real-world scenario, IoT based services improve the domestic environment and are used in various applications. Home automation based IoT is versatile and popular applications. In home automation, all home appliances are networked together and able to operate without human involvement. Home automation gives a significant change in humans life which gives smart operating of home appliances. This motivated us to develop a new solution which controls some home appliances like light, fan, door cartons, energy consumption, and level of the Gas cylinder using various sensors like LM35, IR sensors, LDR module, Node MCU ESP8266, and Arduino UNO. The proposed solution uses the sensor and detects the presence or absence of a human object in the housework accordingly. Our solution also provides information about the energy consumed by the house owner regularly in the form of message. Also, it checks, the level of gas in the gas cylinder if it reaches lesser than the threshold, it automatically books the gas and sends a reference number as a message to the house owner. The proposed solution is deployed and tested for various conditions. Finally, in this paper, the working model of our proposed solution is developed as a prototype and explained as a working model.
Developed a highly efficient mechanism to distinguish the recently introduced DNS-over-HTTPs packets from encrypted web traffic.
Tech Stack: Python(Selenium, Sklearn, Pandas, Tshark, Keras), Docker, Wireshark
GitHubDeveloped a cross-platform mobile application to improve the hyperlocal delivery ecosystem in my hometown during the Covid-19 pandemic.
Tech Stack: React-Native and Firebase.
ExpoIn collaboration with Dr. Apurv Chauhan (Principal Lecturer @ University of Brighton, UK), performed linguistic analysis on tweets related to Covid-19 to understand the psychological impacts of the pandemic across 23 different countries and 5 different languages. A total of 200K unique tweets were analyzed
Tech Stack: Python(Selenium, Pandas, Matplotlib, NLTK) and LIWC
GitHubCreated a Docker Container ideal for computer vision related research and development. The container comes preloaded with all the necessary tools like OpenCV and PyTorch while consuming the least amount of memory
Tech Stack: Docker, Shell, C++
GitHub
