Varun Prasad

Managed the deployment team and headed the deployment of the product successfully across major banks from HK, AU and UK regions. Delivered customer attributes based on the enrichment to help the customers unlock more analytical decision-making capabilities. Built a transaction behaviour breakdown module to understand generic transaction behaviours across different global regions which also helps scale the product to new regions. Built a crowds identification module for an AU based bank to group customers into granular segments for decision making.

Helped in the building and enhancement of a patented module that empowers Banks to leverage their customer data to its fullest potential. The tool enhances transaction level information and adds key missing details that help in 1) Understanding customer behavior in detail like spending pattern, affinity to brands, loyalty etc. 2) Extract Brand, Industry and Location specific behavior from these transactions both of which would have been impossible without the module. Built a model to predict customers who will lookout for Home loans in the upcoming three months using the features extracted from the enriched data for a large UK based bank. This helped them to target potential home owners to get into a mortgage with them and also to roll out tailor made offers to attract these customers.

Built customer propensity model to identify customers who are more probable to upgrade to a premium service that helped the sellers of achieve their revenue targets. Built a demand forecasting pipeline to help the client estimate service consumption across different geographies using a mix of random forests and K-means clustering. Prototyped a profitability prediction framework to try and predict the estimated profit from a technical contract at a very initial stage of the contract so as to set better revenue targets for the year.

Developed an end to end pipeline along with a Sentiment Analysis module to help the Telecom company to segregate negative customer conversation data received from customer care. Identified top trending keywords from customer conversations real time to identify real time customer problems based on the customer care calls and chats. Built a dashboard that that shows real time customer issues and sentiment based on social media posts across different states to enable the client act quickly.

Developed an end to end QC check framework that automates the entire process of QC reducing the time taken to 10% of the manual time taken whilst also providing accurate results. Built an Anomaly detection framework on WPF application that enables user to set custom alert emails driven by an underlying time series model to alert the user of any unforeseen behavior with the refreshed data. Created Tabular cubes of data that fed to Power BI dashboard to help the users consume various metrics required.