April 2020 to June 2020
Integrated a reliable, fault-tolerant emulation of RAFT to achieve consensus-based leader election in GoLang. Significantly improved upon a simple primary-backup KV store, built atop a highly distributed implementation of Map-Reduce.
April 2020 to May 2020
Scrutinized terrorist network in R for hypothesis analysis to model formation and dismantling of terrorist cells. Predicted missing ties, unidentified recruiters via ERGM. Fine-tuned Autologistic Actor Model for actor-based characteristics. Optimized SIENA for longitudinal analysis, based on relational shift, and identified potential (crucial) canaries in the network
Mar 2020 to Apr 2020
Analyzed restaurants from Yelp dataset on cuisine and regional popularity; classified based on health metrics. Developed effective visualizations on Tableau with detailed time-series analysis for restaurants across Chicagoland to potentially guide opening of new restaurants post COVID. Implemented the entire DS pipeline from scratch in Python, and enhanced performance of Machine Learning models by 8.745%
Jan 2020 to Feb 2020
A recipe parser configured to recognize various ingredients, tools and cooking methods used in a given recipe from recipe URLs provided from www.allrecipes.com. We developed a framework to ensure seamless transformations on the recipe along multiple utilitarian dimensions like to and from vegan, scaling the amount to half or double, changing to a favorite style of cuisine, making it healthier, etc. by using our system’s internal representation for ingredients, cooking methods, and tools.
Jan 2020
Used text mining and tweet parsing to identify hosts, awards, nominees, winners and presenters for the “Golden Globes 2020”, along with best/worst/sad/funny moments, red carpet highlights, and snubs, all through sentiment analysis, generalizable for any award ceremony (Oscars).
Oct 2019 to Dec 2019
Constructed NLP system for geographically tagging articles of locally famous people to their birth city and/or alma mater. Created robust NER in Python, managed database server, and designed aesthetic UI for the personalized news feed
Oct 2019 to Nov 2019
Developed models to identify, meaningfully visualize, and flag insincere questions asked on Quora using various ML classification techniques. Performed cross-validation on a dataset of about 56,000 questions tagged as either insincere (1) or sincere (0), with an overall accuracy of 95.075%.
Nov 2019 to Dec 2019
Designed a third-person racing game from scratch in Racket programming language as the Final deliverable for the IPD course. Gained a thorough understanding of language-oriented programming by strictly following the language-agnostic 5-step design recipe for end-to-end industry-standard code design, with complete documentation and near-perfect test coverage.
Mar 2018 to July 2018
Collaborated on a cross-platform (web, mobile and desktop) application for managing and analyzing group as well as personal day-to-day expenses. We did this at a time when a similar application (Splitwise) used to do the same as an Android app but they didn’t yet possess a competitive version for both web and desktop.
Administered databases for a concurrent, synchronized back-end, further intregating it to an easy-on-the-eye design for Web (HTML/CSS/Javascript), Desktop (JAVA) and Android UI/UX with multi-user access support, easing the settling of group expenses, and categorizing/monitoring of personal ones.
May 2018 to July 2018
Managed data-center virtualization on VMWare Workstation cluster formation on MS Server-2008 R2 using SAN. Strategically deployed and tested fail-over cluster (with Quorum) for maintaining cloud-based services and web apps.