Nihit Desai
 |
I am currently studying Computer Science at Stanford University and expect to graduate in June 2017. |
Contact
Experience
Teaching Assistant, Stanford, 2015 - present
Helping teach undergraduate and graduate Computer Science courses
Software Engineer intern, Slack, Summer 2016
Worked on the Data engineering team (Spark, Presto, Hive, Scala)
Software Engineer, LinkedIn, 2013 - 2015
Worked on the Query Understanding and Search Quality team
Research Assistant, UIUC, 2012 - 2013
Data and Information Systems group, advised by Professor Jiawei Han
Education
Stanford University, M.S. Computer Science, 2015 - 2017
University of Illinois at Urbana-Champaign (UIUC), B.S. Computer Science, 2011-2013
Indian Institute of Technology-Madras (IIT-M), B.Tech Computer Science, 2009 - 2011 (transferred out)
Recent projects
Deep Reinforcement Learning to play Space Invaders
Deep Q-Learning architectures to train agents to play Space Inveaders. Course project for CS221 [summary]
Extensions to Tree-Recursive Neural Networks for Natural Language Inference
Improved tree-recursive neural network architectures (Deep Tree-Recursive LSTMs and Attention models over tree-recursive neural networks) for the task of recognizing textual entailment. Course project for CS224d [summary]
End to end system for recognizing and solving Kakuro puzzles
Built a system for recognizing and solving Kakuro puzzles (a category of logic puzzles, often referred to as a mathematical equivalent of crossword). As far as we know, ours is the first attempt to build such a system for Kakuro puzzles. [summary]
Mining Topic Hierarchies from the Microsoft Academic Graph
Independent study project that explores approaches to construct topic hierarchies from the Microsoft Academic Graph, which is a Web-scale entity graph that serves as the backbone of Microsoft Academic Search.
ResFuse and ReFiNet: New CNN Architectures for Image Classification
ResFuse (a modified Deep Residual Network with a projected convolution connecting input and output between two residual blocks) and RefiNet (a CNN-based hierarchical classification framework) are improvements to CNN architectures for image classification. Course project for CS231n [summary]
Plead or Pitch? The Role of Language in Kickstarter Project Success
Predicting the success of crowdfunding campaigns using supervised learning techniques, with a focus on features derived from the language used in project pitches. Course project for CS229 [summary]
Publications
C.Wang, J.Liu, N.Desai, M.Danilevsky, J.Han, Constructing Topical Hierarchies in Heterogeneous Information Networks, KAIS 2015
M.Danilevsky, C.Wang, N.Desai, X.Ren, J.Guo, J.Han, Automatic Construction and Ranking of Topical Keyphrases on Collections of Short Document, SDM 2014
C.Wang, M.Danilevsky, N.Desai, Y.Zhang, P.Nguyen, T. Taula, J.Han, A Phrase Mining Framework for Recursive Construction of a Topical Hierarchy, KDD 2013
C.Wang, M.Danilevsky, J.Liu, N.Desai, H.Ji, J.Han, Constructing Topical Hierarchies in Heterogeneous Information Networks, ICDM 2013 (Best paper candidate)
M.Danilevsky, C.Wang, F.Tao, S.Nguyen, G.Chen, N.Desai, J.Han, AMETHYST: A System for Mining and Exploring Topical Hierarchies in Information Networks, KDD 2013 (system demo paper)
External links: Google Scholar, DBLP