- Building churn prediction service which will enable clientele to discover WHICH of their customers are churning and WHY are they churning. - Prediction service dynamically gets the data, builds several models and saves the best model as part of a cron job. - Whenever the client queries the system for a particular customer, a query is generated and is passed onto the model which allows them to get a prediction of the probability of a customer churning along with the factors that might be causing the customer to churn.- Applying text mining and NLP techniques on support tickets from Zendesk and JIRA along with Notes made by CSM's to get the sentiment along with key phrases to improve response time and tag tickets in order to help CSM's and QA to get a quick idea at what each ticket is about and whether the specified issue is prevalent throughout the platform. Tools being used: Python, Flask, nginx, Salesforce, MongoDB, AWS, APScheduler, Talon, BeautifulSoup, NLTK, LIME

- Gathered insights on health data collected from Apple Watch and Apple Healthkit- Built informative graphs to present performance of patients. Eventually integrated into production app.- Developed a health metric based off of various factors including heart rate and age. (Soon to be copyrighted/patented)Tech Used: Python for data wrangling and pre-processingPower BI - For designing visualizations

• Developed a dashboard to keep track of system up/down times resulting in reduction of down times by 35%• Revamped the legacy attendance system using ASP.NET MVC with SQL Server.• Improved data quality by validating company’s web forms using JQuery and Javascript