Projects & Workshops

• Realtime Voice Call Data Integration, DISH Network, May - July 2023
  Integrated business critical voice call data into the Data Lake. Data was sourced through Kafka from devices on the field, enriched using multiple inventory datasets and persisted in a multi-tenant friendly architecture. It was further integrated into Redshift Serverless in an efficient manner.
• Process Modernization, DISH Network, January - April 2023
  Modernized several processes using outdated libraries and infrastructure to mitigate vulnerabilities and reduce costs. Several short running processes were migrated to serverless architectures which reduced costs by upto 60%.
• Data Pipeline Upgrades, DISH Network, August - November 2022
  Upgraded several data pipelines and an Apache Druid cluster to support a massive increase in data volume due to the expansion of Dish's 5G Network. Capacity was increased from 20-30 pipelines bringing 10 GB data per day to more than 200 pipelines bringing 3 TB data per day. With a requirement to be available 24/7 and very low error rate.
• Geospatial Network Analytics, DISH Network, March - July 2022
  • Correlating data from UE, RAN and Transport domains for simple but effective network monitoring.
• Jira Ticket Automation, DISH Network, July - October 2021
  • A ticket automation pipeline that lead to faster time to resolution.
  • It is a continuously learning NLP and machine learning pipeline that assesses the quality of tickets and recommends assignments with 82% accuracy.
  • Makes use of Amazon Comprehend, Lambda, DynamoDB, Parameter Store and S3.
• Network Analytics Application, DISH Network, January - March 2021
  • A web application with elements of service assurance, customer assurance and social media analytics.
  • Built KPIs and visualized them with Apache Superset for the network data for the 5G Wireless program.
  • Built KPIs and developed custom visualization widgets with D3 and Plotly for social media analytics data for the 5G Wireless program.
• SatelliteNet: Satellite Image Smoke Detection, GWU, Spring 2020
  • Detection of smoke in satellite images using deep learning.
  • Improved accuracy to 94% from 92.75% and Cohen's Kappa to 0.9267 from 0.9130 over previous models.
  • Implemented with easy to use training and evaluation scripts using PyTorch.
• Analysis of Mixed Quote Representation in News Sources, Natural Language Processing Project, GWU, Fall 2019
  • Quantitative analysis of partial representation of full quotes.
  • Web scraping for data collection.
  • Analysis using Jaccard similarity and k-means clustering.
• American Sign Language Recognition using CNN, GWU, Fall 2019
  • Recognize basic alphabet hand signs from images.
  • Use pre-trained and fine tuned ResNet model from PyTorch.
• Bayesian Analysis of Play Store App Ratings, GWU, Fall 2019
  • Hierarchical modelling of app ratings based on installs, reviews and price.
  • Generalized Logistic Regression model created using Markov Chain Monte Carlo (MCMC) in JAGS.
• Earthquake Time Prediction, Machine Learning Project, GWU, Spring 2019
  • Entry for the LANL Earthquake Prediction competition.
  • Trained models including linear regression, elastic net and neural networks to predict earthquake times using real time seismic data.
  • Model selection and hyperparameter tuning using random forest.
• Traffic Simulation with Cellular Automata, HPC and Parallel Programming Project, GWU, Spring 2019
  • Two lane traffic simulation using cellular automata.
  • Implements the algorithm described in this paper.
  • Parallelised using OpenMP in C.
• Olympians, Data Visualisation Project, GWU, Spring 2019
  • Data exploration tool for the summer Olympic games.
  • Uses d3 for visualisation.
• Prediction of Kickstarter Campaign Success, Data Mining Project, GWU, Fall 2018
  • Analyzed a Kickstarter campaigns dataset.
  • Built a prediction model for success using logistic regression based on catergory of the project, country of origin, duration of campaign and number of backers.
• Prediction of a country's performance in Olympics, Data Science Group Project, GWU, Fall 2018
  • Built a linear regression prediction model for proportion of medals won by a country based on economic factors like GDP, Population, host advantage as found in a previous project and performance in the previous Olympics.
• Does the host country of Olympics have an advantage? Data Science Group Project, GWU, Fall 2018
  • Analyzed data from summer Olympics to determine whether countries win more than their average when they’re hosting Olympics using a t-test.
• Interaction of Catalytic Active Motors with Surface, Final Year Project, IISER Bhopal, 2016 - 17
  • Worked on developing a numerical simulation in C.
  • Simulation consists of a system of 10,000 particles interacting with each other and a wall through Lennard-Jones potential.
  • Involved grid search for nearest neighbors, periodic boundary conditions and application of Anderson thermostat.
• XX SERC Preparatory School in Theoretical High Energy Physics, IISER Bhopal, July 2015
• Application of Bounded Bhattacharyya Distance for Gravitational Wave Detection, IAS Summer Research Fellowship, at IIT Gandhinagar, Summer 2014
  • Developed application of BBD to signals.
  • Found out how BBD differs from other distance measures like KL divergence.
  • A New Family of Bounded Divergence Measures and Application to Signal Detection Shivakumar Jolad, Ahmed Roman, Mahesh C. Shastry, Mihir Gadgil, Ayanendranath Basu, Proceedings of the 5th International Conference on Pattern Recognition Applications and Methods - Volume 1: ICPRAM, ISBN 978-989-758-173-1, pages 72-83. DOI: 10.5220/0005695200720083