India’s First Bank Led Datathon

India’s First Bank Led Datathon 1

One must be familiar with ‘hackathons’ and similar code challenges conducted by a plethora of different tech giants to excavate the hidden talents and throw them the business use cases to solve tiny segmented bottlenecks. What you might not be familiar with is the actual time line we’re looking at here. Any such model or product developed in a hackathon gets prototyping time of 24 hours or so. But the production time is a hefty bet.

YES Bank, a premiere name in the Indian banking industry, operating as a corporate bank, with retail banking and asset management as subsidiary functions, saw the bigger picture. The fourth largest private sector bank in the country sought to involve the data science community outside of their organization and see if their use cases and ingenuity could solve some of the greatest but unforeseen setbacks in the financial system.

Datathon, launched in September 2018 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. Over the last 70 days, the participating got the opportunity to work on YES Bank’s anonymized and curated data sets to bring forth their own models and solutions to the problem.

The organization reached out to the global developer, coder and data science community to work on YES BANK’s data stacks – using curated, tokenized and anonymized data. The program attracted 5500+ data scientist/engineers and developers who participated in 4 diverse Machine learning challenges over a week’s time, to qualify for the main challenge. 200 top data scientists identified from the challenges were given 60 days to create working data models/prototypes which will be trained, tested and deployed by the bank by January, 2019, thus, reducing the time usually taken to bring groundbreaking models into production by almost 300%.

The top 200 teams include not only students from top technology institutes like IIT Bombay, Chennai, Kharagpur and ISI Kolkata, but also 150+ professionals from organizations like IBM, Walmart Labs, Siemens, Amazon Development Centre, Capillary Technology, TCS, Accenture, Amdocs and Infosys among others, who have formed 30 different teams working round the clock across the weekend to put forth the most unique model. What’s more enthralling, is the fact that these data professionals from a non- financial services background are participating in the challenge beyond the work hours to make a difference.

The Amazon Development Centre says, “Financial services industry is fairly unique in terms of use cases and problem statements that data analytics professionals face. For us Datathon is a first of its kind opportunity where we are working almost as an extension of the bank’s data analytics team, creating live data models on practical use cases and data stacks”.

The Bank has also onboarded a pool of 20+ data science experts and leaders across industries, to support the data science teams as well as help the bank identify and share use cases for product development using Big Data Analytics, ML and AI. The experts from data driven companies like Cloudera, Amazon Web Services, among others will work on an ongoing basis with the Datathon team.

“The participants come from diverse backgrounds, like there are about 8-10 teams who are from start-ups. And of course there are students from NITs, IITs including a college from Sri Lanka. This gives variety of ideas and use cases with a surplus of heterogeneity of approaches. And therefore, our team of data science experts work very closely with the teams as they identify a use case and suggest working on a model based on that data set. This is important, as we, knowing the industry, are able to give a ruling as to which use cases stands on priority. The teams and participants here have shown a zeal to work towards solving problems, developing models and watching their models put into action, rather than securing the prize money”, mentioned a senior representative from the bank’s Fintech and strategy team.

Catching up with the Participants- Excerpts

Crunching planet sized data throughout the program, the teams have come up with astonishingly impressive models and use cases. A few words with some of the participating teams reveal the level of understanding that these participants have acquired of the industry’s real pain points. Anand Krishnamurthy, working for TCS says, “It’s a whole new different world out here, and it’s pretty cool with the sheer size of data (AWS) that we are getting to work with. As for my use case, I am working out a model that integrates, categorizes, prioritizes and identifies the key phrase/words from the tickets raised by customers”.

Akshar has been working on multiple use case models and he explains, “Our primary objective is to impact the customer experience and revenue stream for the banking entity. At the customer front, we intend to identify their transaction patterns and provide them with reminders to help them stay on track with their monthly commitments. For the bank, we use this same transaction pattern in a different form of fraud detection. A transaction pattern deviation from 50k on an average to a whopping 2Lacs might raise eyebrows but a series of 20 transactions of 10k might just slip out. We are building a model where such unique scenarios are detected and notified to the bank. On this journey, we realized that this is not just any other coding event. We’re not just coding but also analyzing the business impact of our codes. The guidance, exposure and learning experience from our fellow corporate participants is great.”

Jalpesh and Salim from Vodafone and PwC are excited about their, “We are working on model that helps enhance the RM productivity. In far more often cases, RM doesn’t usually get through potential customers due to lack of visibility affecting the overall revenue and customer base of the bank. What we’re doing is, based on an analytics engine, we have developed 6 models. These models will assist an RM to understand which customer to approach, when to approach, what products to pitch in etc. It’s a great opportunity for us and we look forward to help the industry by nd large and make a difference”.

The program featured 15 products on 23rd December (Sunday) at Bangalore’s RMZ Infinity chambers and after evaluating will be taking few of them into production immediately. January will witness new dawn in the banking system for YES Bank. Hopefully, the entire banking industry will follow in the year ahead.



Written by Srikanth

Passionate Tech Blogger on Emerging Technologies, which brings revolutionary changes to the People life.., Interested to explore latest Gadgets, Saas Programs

Design Thinking

Novel AI can detect age, gender of people on videos

Decentralized AI Network SingularityNET Launches Developer Portal 2

Decentralized AI Network SingularityNET Launches Developer Portal