Boston Consulting Group (AKA BCG) is a Global Consulting Firm that partners with leaders in business and society to tackle their most important challenges and capture their greatest opportunities. By combining advanced technology with deep industry experience, BCG GAMMA creates custom solutions for the artificial intelligence generation that bring about end-to-end business transformation. I worked on Retail Catalyst product which helps clients strategize pricing and facilitate simulation runs.

I was working as a Software Engineer Intern. We had Python in Backend and for the Frontend JavaScript, HTML, CSS were used. We were migrating from Flask-MVC Framework to Django-FAST API framework & from Vanilla JS to Angular JS. Majorly we used to work on the new features or the bugs we have identified. I developed new features like integration with Slack where only error messages will be posted, which makes the debugging easier for the developers. Also, Enhanced the robustness of the application by resolving the identified issues and minimized the frequency of error messages.

Worked on this project under prof. Tomasz Imielinski, where users can generate synthetic dataset. Where users can provide the attributes & rules based on which they want to genearate dataset and application will generate data in just few clicks. For class Data101 at Rutgers, professors used this tool to generate quizes as well. It helped them save their ~80% time. Users can also provide the foriegn data (e.g. data imported from Kaggle) and tweak the data distribution to make more interesting patterns. One of the quizes can be found here .

I created PRD & developed the application from scratch. I worked as Full Stack Developer. Backend was written in Python and for the Frontend JavaScript, HTML, CSS were used. In order to generate random data, I used Random Forest and Decision Trees algorithms. For the version control and deployments, we used GitHub and AWS EC2. Released the first version of the application within 5 months.

When I was looking for the Internship/ Co-op/ Full Time opportunities, I used to apply to probably hundreds of jobs in each recruitment season and often forget to keep track of them. You too? Then this application would be very much useful for both of us! It helps you stay organized.

There are very famous job boards like Jobs.lever, greenhouse.io, Workday. Most of the companies use one of these. Wrote web scraping script to fetch all the details regarding this job from URL itself. You don't have to fill each detail manually. Phew! Also, you can keep track of which resume did you use for this job. You can add, delete and search through your job apps. Lastly, you can change the status of jobs which will prevent you from filling out the same application twice.

Since me and my classmates were facing this problem, I reached out to them to understand what all features they would like to have in the application. I designed the prototype & developed the application to address the issues. I solely worked on this project. Backend was written in Java (Spring-Hibernate framework) & for the webscrapping I used Python. For the Frontend JavaScript, HTML, CSS were used. Used MySQL as database. For the version control and deployments, I used GitHub and AWS EC2.

When I learned Item-Item Collaborative Filtering, I wanted to apply it to day-to-day tasks. Hence, developed web application where user can play songs and get recommendations based on the preferences. Used subset of data released by Echonest as part of The Million Song Dataset Challenge. It considers different measures of songs like acousticness, liveness, speechiness, loudness etc. Implemented K-NN & Item-Item CF algorithms to find similar songs from the dataset and recommend those to user. Also, users can play all the songs (from the dataset) in the application. Since, I didn't have audio files for any of the songs, scrapped YouTube in order to get MP4 file and then converted it back to MP3 on-demand.

I solely worked on this project. Backend was written in Java (Spring-Hibernate framework) & for the webscrapping & recommendation algorithms Python was used. For the Frontend JavaScript, HTML, CSS were used. Used MongoDB database. For the version control and deployments, I used GitHub and AWS EC2.

During my 1^st semester here at Rutgers, we were asked to implement A* algorithm in the toy example. But I wanted to see how it works in the real world data. So, I used it in order to find shortest path from given any source and any destination. Used OSM in order to extract the data in the graph & edge format. Which had ~27k nodes and ~63k edges. It always tries to find the optimal path. So, in most cases it shows the path which is suggested by Google when you select travelling mode as "Walk".