Tools/Languages used :

Python, NLP toolkit, Hadoop, Amazon Web Services

Descripion :

Implementaion of complete data science pipeline.

Data Cleaning

Data Integration

Data Prediction

Data Vizualization

Space of product reviews is unstructured. User has to read lot of reviews to get complete and unbiased view of the product. This project aims to solve this problem using Data Science.

Complete Report of project can be downloaded here.

Tools/Languages used :

Scala, Spray-Can, Spray-Json

Descripion :

REST based Facebook Api server implementation using Actor model. Data handling using CRUD. Security is implemented using end-to-end encryption and Digital Signatures.

Complete API documentation

Tools/Languages used :

Hadoop, Java, Amazon Web Services

Descripion :

Google works on Page Rank algorithm. Google indexes pages and display search results on basis of Page Rank algorithm. We used the algorithm to determine rank of each Wikipedia page.

Implementation :

Our source dataset is huge Wikipedia corpus. Data cleaning included extarcting title names of each wikipedia page and outlinks of the same page. Second step included removing redlinks from the extracted data. Page Rank algorithm is implemented on these extracted data for eight iterations. Finally we get weight of each wikipedia page.

Everything runs on Amazon Web services.

Tools/Languages used :

Python, OpenCV

Descripion :

I like solving sudoku puzzles in daily newspaper but hate to wait for soultions till next day. Why not make something with which I can verify solved sudoku results in seconds. This computer application detects sudoku puzzle from camera and solves in fraction of seconds.

Implementation :

Implementation involved using Computer Vision(to detect sudoku grid), Machine Learning (to detect digits in grid) and Sudoku solving algorithm. Using contours in computer vision, big grid and small nine grids are identified. KNN trains 0-9 digits from training data which is used to detect digits. Digits with its positions are feeded to Sudoku Solving algorithm.

Tools/Languages used :

Python, OpenCV

Descripion :

This published paper is result of my undergraduate research project. It presents a simple approach to detect digits from images. It presents a simple OCR algorithm for digit detection.

You can pretty much read the paper itself here.

Tools/Languages used :

Scala(Actor model)

Descripion :

This paper uses consistent hashing and finger tables to find and query a node in logarithmic time in distributed systems.

Chord: A Scalable Peer-to-peer Lookup Service for Internet Applications

Tools/Languages used :

Scala(Actor model)

Descripion :

This paper uses push sum protocol to compute average over large number of nodes in distributed systems. Communication happens over Gossip protocol.

Gossip-Based Computation of Aggregate Information

Tools/Languages used :

Matlab, Qt, C

Descripion :

This project was undertaken as a Software Engineer at Tata Consultancy Services. Our team developed HMI Interface of Sync 3.
Software development cycle consists

Modelling the required functionality(Matlab)

Converting models to C or C-mex code

Developing User Interface in Qt(C++ and QML)

Writing network interface(in C) to connect models with User Interface

My Part :

Our team was of four people. Everbody was involved in each task. Though I was more focused in developing User Interface and Network Interface as mentioned above.

Tools/Languages used :

Powershell, ShowUI

Descripion :

This project was undertaken as a Software Engineer Intern at Tata Consultancy Services. We developed network management software(Administration helper) in Powershell from Scratch. Some of the features of developed software were :

Managing a group of computers remotely using GUI developed(using ShowUI)

Taking remote access of a single computer or group of computers and running bunch of commands

Single click mass installation of most used softwares

Single click Update and Maintenance of large number of computers

Tools/Languages used :

Python, Plotly

Descripion :

This is actually an entry to Data Visualization contest on Instructable.com. It was mandatory to use Plotly api for visualizing data.The code here scans all your posts on timeline and displays a chart based on which friends have highest likes on your posts.

I would be quite happy if you try it out yourself. Here is my Instructable.

Tools/Languages used :

Python, Latex

Descripion :

This small project is result of necessity and boredom. Our undergraduate advisor needed a journal having all lab codes I have used in my semester along with their description. All codes were stored in .m(Matlab) files. It required tiresome work of few hours copy pasting everything into word file. Instead I just used my new learned Latex skills and python to automatically create journal and save lots of time and energy. Infact all of my colleagues found it useful and used it to generate their journals.

Do checkout code. I have uploaded the code in my Github Repo.