Technical Skills


Programming Languages

  • Proficient: C, C++, Java, Python
  • Good: HTML, CSS, Javascript, PHP, Shell Scripting
  • Beginner: ARM Assembly, R, Matlab, Ruby

Tools Used

  • Operating System: Mac OS X ElCapitan, Linux (Ubuntu 14.04 LTE), Windows 8
  • Database Systems: MySQL, PostgreSQL, Oracle, MongoDB
  • Text-Editor: Sublime Text, Vim
  • Development Tools: Git, Latex, GDB

Areas of Interest

I spend most of my time reading about web technologies and experimenting with them. My main interest lies in the areas of artificial intelligence and machine learning.


Libraries

  • Web: Spring, Django, Ruby on Rails, Node JS
  • Desktop: JavaFX, Swing
  • Frameworks and Libraries: Keras, DeepLearning4J, Terrier,Mallet, Apache Tika, Apache OpenNLP, Apache Lucene
  • Others: Guava, Jsoup, CRF++, CRFSuite


Projects


Olympedia - Olympics Information Base

  • Implemented Scripts to extract data from wikidata.org and populate MongoDB database
  • Implemented module to display players' personal information.
  • Implemented module to display country's performance over a range of years. (Used Morris JS to do visulualisation)
  • GitHub Link: here.

PDTB-styled Shallow Discourse Parsing

  • Implemented a Shallow Discourse Parser based on Lin et.al. pipeline.
  • It returns all discourse relations in a text in the form of discourse connectives, its two arguments and the relation sense.
  • Achieved an F1 score of 0.11 on blind dataset provided in CONLL-16.
  • Github Link: here.
  • Task Paper: here.

FIRE-2015 Shared Task on Mixed Script Information Retreival

  • Subtask 1:
    • Identify language of a word in code mixed data.
    • Trained CRF classifier using Linguistic and contextual features.
    • Achieved a weighted F-score of 0.804.
  • Subtask 2:
    • Retrieve mixed-script documents from a corpus for a given mixed-script query.
    • Used DFR to retrieve documents.
    • Achieved a NDCG@10 score of 0.7160.
  • Github Link: here.
  • Task Paper: here.

CLEF-2015 eHealth Named Entity Recognition for Clinical Data

  • Identify named entities in a French medical document.
  • Trained CRF classifier using linguistic, contextual and MetaMap thesaurus features.
  • Achieved an F1 score of 0.67.
  • Task Paper: here.

Amazon Home Services:
Seller Image Cropping Map

  • Developed a tool for enabling sellers to crop iages before uploading.
  • Wrote Unit tests and Integration tests to verify the code behaviour.
  • Was able to push code to Gamma stage.

Amazon Home Services:
Seller Store Map

  • Used Nokia HERE Map API to display location of any seller store.
  • Used iFrames to prevent API access to customer and seller private information.
  • Wrote Unit tests and Integration tests to verify the code behaviour.
  • Was able to push code to Beta stage.

Hospital Management System

  • Led a team of 9 people to implement a desktop software for Sharda Clinic in Gandhinagar as part of Software Engineering course.
  • Features included: Prescription Creation, Financial Management, Test-Report Maintainence, Inventory Management and Medication Report generation for indoor as well as Outdoor patient.
  • Was responsible for client communication, designing database and implementing prescription creation feature.
  • Github Link: here.

File Sharing Desktop App

  • Implemented a filesharing multi threaded desktop app using socket programming.
  • Implemented feature to download same file from multiple peers using round robin scheme.
  • Github Link: here.

Robocup: Multi-Agent Soccerbot

  • Feature extraction module was implemented to obtain the state of each player.
  • Q-learning module was implemented to learn rewards for each state transition.
  • ε-greedy method was also implemented to maintain a balance between exploration and exploitation.
  • Github Link: here.

Learning Management System

  • Implemented the Learning System prototype along with recommendation engine
  • Generated a concept map based on students' answers to questions
  • Recommend courses taken by the students' peers (Collaborative Filtering)
  • Recommend courses in which his knowledge level is decreasing (Remediation)
  • Github Link: here

Dynamic Memory Allocator

  • Used explicit free list and next-fit algorithm to optimise the search of free blocks in memory.
  • Used caching to allocate heap memory to frequent requests of similar sizes.
  • Github Link: here.

Discourse Annotation Tool

  • Implemented a tool for annotating discourse relations in text.
  • Features:
    • Select portion of text for connective and argument spans.
    • Select sense of relation (PDTB sense hierarchy).
    • Upload new documents for annotation (admin access).
  • Website: here.