Advisor: Professor Lise Getoor Research area: I am interested in machine learning and probabilistic graphical models in general, and applying these to various user behavior modeling tasks. My research focused on 'Statistical Relational Sentiment Analysis', i.e, modeling 'sentiment' in various relational domains(online reviews, MOOCs, social networks) using statistical relational learning techniques, in relational and network data.

Research : Statistical Relational Sentiment AnalysisBuilding unsupervised machine learning models, to identify 'attitudes' between organizations based on the hierarchy and relational structure in the organization network.Aspect-based sentiment analysis in online course discussion forums : Developing a weakly supervised joint model for aspect-sentiment in in MOOC forums, by modeling the dependencies between various aspects and sentiments.Fine-grained sentiment analysis:Modeling sentiment at a fine-grained or phrase level, in online reviews, using relational structure in text. Course Projects : Twitter Sentiment AnalysisThis project involved applying machine learning techniques to classify the tweets as a positive or negative sentiment.Github Link: https://github.com/shachi04/TwitterSAHaskell implementation of Document Search and Ranking based on Word relevance -Implemented a library for finding tf-idf of words in documents, and also similarity between documents using cosine similarity measure. - Developed a web application using the Warp web server for Document Search and Ranking based on word relevance using tf-idf measure. - Github link : https://github.com/shachi04/HaskellTextAnalysis

Course : Algorithms and Abstract Data TypesResponsibilities include : Teaching during weekly section hours, helping students with examples, programming questions and grading exams. Topics taught : Complexity analysis, data structures - linked lists, priority queues, dictionaries, heaps, balanced trees, hashing and graphs, sorting and searching algorithms.

Worked on identifying "intent" or action in SMS conversations using probabilistic models and NLP techniques.

-Involved in the design and implementation of Gateway features to integrate the telepresence products of HP and Polycom.

- Worked on the development of an analytic application of Halo involving usage data, by developing reports that provide information to the Halo customers about meeting duration, the duration concierge takes to attend a call, studio downtime, and so on.-Worked on the testing and maintenance of a Gateway application to integrate HP Halo and Microsoft Lync at the partner site in Corvallis, Oregon, USA.

Implemented a software solution through which a hardware device is emulated and virtualized, thereby eliminating the need to have the hardware device to be present physically on the latter computer.