• Machine Learning Modeling: Enabled Stanford researchers to analyze the textual data of U.S. Securities and Exchange Commission (SEC) filings along with numerical data by creating a corpus-wide set of word embeddings using Word2Vec models in PyTorch and also expanded the financial sentiment dictionaries using NLP techniques which improved regressions on stock return volatility significantly.• Data Processing and Analysis: Developed text processing and parsing pipelines for analyzing various types of SEC documents using Python and analyzed statistics of parsed documents over the last 10 years.• Evaluation and A/B Testing: Benchmarked a variety of novel approaches using techniques like bootstrap, tf-idf and multiple A/B testing for word embeddings evaluation; Fine-tuned BERT model for the classification of certain vs uncertain statements in financial documents and achieved the weighted F1-score of 0.93.

• Designed and developed interactive visualization and reporting dashboards that monitored growth KPIs for a global beverages company which led to time and effort savings of two person days per week and also introduced a company-wide single source of truth for growth KPIs, which were updated in near real-time.• Project included Excel upload functionality of departmental data to store into the database (SQL server) and provided data visualization of key metrics on various dashboards utilizing REST APIs, Angular, D3.js, Python and generated reports automatically and on-demand.

• Performed customer data analysis to derive business decisions and data monitoring using CRM data for Adidas. Created data reports using Tableau and customized various dashboards in Grafana (open source visualization and analytics software). • Developed and deployed a new technology stack upgrade in France Telecom for call data records (CDR) processing pipelines which reduced processing time from 5 hours to ~8 minutes.• Developed collection module for the project in Java and UNIX which parsed the Call Data Record files in BCD format and then inserted it into the database and processing module responsible for applying the business and validation logic in PL/SQL.• Lead infra requirements gathering and release activities of the project.• Customized existing plugins in Grafana like Diagram Panel, Status Panel, World Map Panel which are built using Leaflet, Mermaid JS as per client requirements and created new panels following Grafana architecture.

• Developed responsive web applications including OverTheTop (OTT) streaming platform and Customer Sentiment Analysis Application using JavaScript frameworks (Angular, Backbone and Node JS), NoSQL(MongoDB and CouchDB) and IBM Cloud Services such as IBM Watson, IBM SingleSignOn.• Created efficient web interfaces using HTML5 (new tags like audio, video, canvas), CSS3, Bootstrap, JavaScript, jQuery, AJAX, JSON and developed several REST web services.• Provided AAA (Authorization and Authentication) capability to a customer sentiment analysis application using IBM Single Sign On service and integrated it with the backend in NodeJS.

• Courses Taken: Object Oriented Concepts using Java, RDBMS, Javascript 
and J2EE framework (EJB, JPA, JSF).

• Developed College Student Management System using the .Net framework and SQL database which enabled storage and access of student profiles, courses taken and marks obtained by each student.