YES BANK conducted the finale of YES Datathon launched in September 2018 on December 22-23 in Bangalore. This was India’s first bank led Datathon aimed at augmenting YES BANK’s embedded data analytics & ML units to drive rapid prototyping of AI/ML based products, optimize digital product suite, and enhance product/service design and delivery, in an accelerated 100 day period.
The initiative saw participation from over 6000 data scientists/engineers and developers who participated in diverse Machine Learning challenges. Some of the top models/products developed were based on:
- Predictive service delivery to enable the bank to provide proactive service
- Sales recommendation models
- Customer Transaction Relation Graph, equivalent to a social network of transactions
- Fraud Detection and analytics for card customer
- Customer Sentiment Analysis to predict Customer engagement etc.
- Customer 360 tool to provide comprehensive products & services to each customer
- Personal Finance Management tool without requiring access to customer SMS
The top 200 teams comprised of students from top technology institutes like IIT Bombay, Chennai, Kharagpur and ISI Kolkata along with professionals from organizations like IBM, Walmart Labs, Siemens and Amazon Development Centre among others. The teams received support from over 20 data science experts from across industries who helped them in the product development process.
At the finale, the top 50 teams showcased their final models which were trained and tested on more than a terabyte of created and tokenized datasets – the 10 best solutions won cash prizes worth Rs.20 lakh. The winning team Oracle was adjudged ‘Data Wizards’, for creating a ‘360 degree model for profiling of every customer to predict next best action’. Team Insight from Sri Lanka was adjudged the best student team and was also one of the top 5 teams. This student team developed a model for ‘predicting customer product and service affinity’. The other teams in top 5 were Finance Data Dons, Reverse Atlas and Django Unchained, who presented models on Predictive Payments to create a Personal Finance and Investment Management Tool + POS Anomaly Detection, Creating a community akin to social media sites using transaction analysis and Predictive model for customer relationship managers to target existing and new customers using transaction nodes, respectively.
The top 20 models identified by the Bank will be taken live within a month and the remaining will be moved to the Bank’s product library to be iteratively developed and taken live within a period of three months. This will also be a first of its kind data product library which will be iteratively expanded to 100+ products within 6 months
Talking about the event, Rajat Monga, Senior Group President, Financial Markets, YES BANK, said “YES BANK embarked on a data centered business model as part of our TechTonic initiative and now has a full stack of technology and talent capability built up. In order to leap frog on this data native transformation, YES Datathon provides us with an opportunity to engage with 6000+ data scientists. It has helped us identify newer use cases as well as statistical techniques and also incorporate cross-industry best practices.
Going forward, YES BANK will also host AI/ML challenges and data engineering workshops to deepen practical and technical knowhow of future technology leaders and to facilitate this, has partnered with top IITs and BITS Bombay to further develop the data science ecosystem, allowing students the opportunity to build algorithms and data models in a deployment ready environment. The 6 Machine Learning challenges hosted in partnership with the IITs and BITS will lead up to Datathon 2019 to be hosted in June 2019.