About Me

Hello! Welcome to my website! I am a graduate student at University of Rochester. I like to build robust softwares. In general I strive to better understand how human interact with computers/machines/platforms and create fun and effective systems that makes computing seamless to interact with and adds more insights to our understanding of how technology impacts our daily life.

I am advised by Professor Henry Kautz. My research interest is in the cross-section of social computing and Human Computer Interaction. Before coming to University of Rochester I worked at KAYAK, Noodle (a search engine for education based services) and Offerpop (a social media marketing platforms). I did my undergrad at Bard College, majoring in Computer Science.

I am pretty outdorsy. Love aything that involves any kind of sports. Soccer, tennis, squash, volleyball, basketball, you name it... Of course I am pretty geeky.

Summer 2016: This past summer I was at eBay working on Big Data problem for the Shopping Experience Applied Research group at eBay. In order to improve the user experience millions of Structured Search Result Pages are generated by eBay. The objective of my project was to detect such similar, duplicate and near duplicate pages; and train a machine learning model to detect pages that may look similar. This helps eBay avoid creating thin content, add value to Structured Data, etc

Below you can take a look at some selected projects of mine and I invite you to check out my GitHub if you're interested in seeing more in detail.


Some Fun stuffs!

  • Multitouch

    Inspired by one of apple's games named, “Osmosis” we thought it would be fun to have that particular application on a multi-touch coffee table surface. It is multiplayer game that incorporates some form of augmented reality part to this platform. Using IR LEDs and Kinnect's IR camera we detect fingers touching the screen; and track moving objects in the game based on the finger movements.Some of the major challenges that we have to overcome are: building the physical surface and syncing the camera and projector, tracking multiple objects and finally coding the game.

  • Sentence Similarity

    English sentence similarity measure is used in a vast number of applications such as online web page information retrieval systems, online advertisements, question answering dialogue systems, text summarization, text mining. Over the years, a number of algorithms have been proposed for this difficult problem, but none of the proposed algorithms give sufficient good answer.
    In this project, we explore three different algorithms for computing English sentence similarity. The first algorithm, which is well-explored in the literature, weights words in each sentence according to term frequency and inverse document frequency (tf-idf ) and uses no semantic information. The second algorithm uses measures of the semantic distance between words belonging to the same part of speech. The third algorithm combines the tf-idf scores and the semantic distance scores between words.
    We evaluate the performance of the second and third algorithms on two data sets: O’Shea’s set of sentence pairs with human similarity judgements [Li et al., Aug, Rubenstein and Goodenough, 1965], and Microsoft Research’s sentence-level paraphrase dataset [Rus et al., 2012]. On O’Shea’s data set, the third algorithm more accurately matches human judgments than the second. On the Microsoft data set, there was not a significant difference between the two algorithms.

  • The IMP

    The IMP (Intelligent Mobile Projector) is an experiment in novel forms of mixed-reality and human robot interaction by combining recent advances in mobile robot and projector-camera systems research.
    Adding mobility to a projector-camera system enables a new class of mixed-reality interfaces. For example, how can we use autonomy as a key ingredient in augmented reality applications? Or can we use controlled mobility to create large expansive surfaces of interaction? Likewise, adding a projector-camera system to a mobile robot adds a new dimension for human-robot interaction. The IMP is able to project almost anywhere using its pan-tilt-move interface.

  • Jajantor-Mamantor

    Jajantor-Mamantor is a virtual maze exploration game. The idea is that the servercontains the main map of the maze and the clients explore the maze remotely/ locally. The game starts from a certain point and everyone’s target is to solve the maze and reach the target position. Every client requests for map information from the server where the entire map is available. This means that the server is continuously looking for data request and responds to the corresponding request.

  • Cheap Explorer

    Encouraged by the idea of how Mars Rover wanders around and collect data and sent them back to Earth, we wanted to build a system that would incorporate bunch of cheap robots exploring the real world and build an online map of the world as they wonder until they manage to reach the target goal/s. We want to build a system that would enable a robot to cover every point in the workable area and find the most optimum path to the goal. In order to achieve this we have implemented the online multi-robot coverage algorithm proposed by Kaminka in the “The Giving Tree: Constructing Trees for Efficient Offline and Online Multi-Robot Coverage” paper.