Skip to main content
  • English
    Discover SWIFT
  • Español
    Descubra nuestros contenidos en español
  • Français
    Découvrez notre contenu disponible en français
  • 中文
    了解我们提供的中文内容
  • 日本語
    日本で入手可能なコンテンツをお探しください
This content is not available in the selected language
Swift Hackathon 2021: The results are in!

Swift Hackathon 2021: The results are in!

Sibos,
14 October 2021 | 4 min read

After much thought and deliberation, the judges have made up their minds. Read why Deutsche Bank and J. P. Morgan have been crowned the winners of the Swift Hackathon 2021.

Over the last eight days, the Swift Hackathon judges have been hard at work – putting their heads together to decide which teams should be crowned the winners of this year’s competition. But it’s been no small feat, with 25 teams and 180 participants to choose from and no shortage of bright ideas. 

“Being a part of this competition has been fantastic,” says Christian Kothe, who’s sat on the judge’s panel for this year’s event. “After careful consideration, we’re delighted to announce that the winners are Deutsche Bank and J. P. Morgan for challenges one and two respectively. I’d also like to extend my congratulations to both of our runners up – Tata Consultancy Services for challenge one and Exactpro for challenge two.” 

Challenge one winners: Deutsche Bank

To be in with a chance of developing the winning solution for challenge one, Deutsche Bank’s team of nine needed to create novel data representations that could be used to better detect transaction anomalies. The need to achieve this has never been more pressing, with research suggesting that 95% of triggered alerts are actually found to be false positives.

Despite initial challenges with the synthetic data, Deutsche Bank’s diverse team from around the world soon found their groove in producing an effective solution. Their strategy was to split into smaller groups and carefully test different approaches, before uniting on the most promising options in a final push to claim victory. 

“After performing an explorative analysis of the data, we were able to understand the underlying patterns and use our business knowledge to engineer meaningful features,” says Lukas Boschert, a data scientist within Deutsche Bank’s Corporate Bank Technology function. “This was critical in the development of an AI solution that could successfully identify the majority of fraudulent transactions. This could have tangible benefits in the real world – adding an additional layer of security for compliance teams around the world.”

This could have tangible benefits in the real world – adding an additional layer of security for compliance teams around the world.

Lukas Boschert, Data Scientist, Deutsche Bank

Challenge two winners: J. P. Morgan

J. P. Morgan’s team totalled eight members, all working within the bank’s specialised innovation and blockchain unit – Onyx. The challenge saw them go head-to-head with other teams in a bid to create synthetic data sets that can be used to train machine learning (ML) services, all while maintaining customer privacy. Banks and other industry players are already starting to use these intelligent algorithms to recognise transaction types, learn from user behaviour and help hone the customer experience. 

“First, our team had to properly understand the process that generates real data – the trends and patterns that are typically seen in fraudulent transactions,” says Monik R Behera, a software engineer within the Onyx engineering team. “Once we’d done that we could begin modelling that process to create fully anonymised synthetic data sets. We began the Hackathon with three or four algorithms that we needed to test, and worked to narrow these down into our winning solution.” 

Our success here means a lot more than just winning a competition. It has the potential to provide the financial community with the data they need, regardless of their size or resources, and collectively work to reduce fraud throughout our industry.

Ker Fran Lee, Product Manager, J. P. Morgan

“Today, not all institutions have equal access to data sets that allow them to train ML models and subsequently, detect fraud,” continues Ker Fran Lee, who works as one of Onyx’s product managers. “Our success here means a lot more than just winning a competition. It has the potential to provide the financial community with the data they need, regardless of their size or resources, and collectively work to reduce fraud throughout our industry.”

The journey has only just begun

While the Hackathon is a great opportunity to bring innovators together, kick start the creative process and accelerate product development, innovation shouldn’t be limited to one event a year. Whether you’re a bank, fintech, corporate or vendor we want to hear from you, even if you didn’t manage to take part in this year’s event. 

“We’re encouraging our community to come and innovate with Swift,” says Tom Zschach, Chief Innovation Officer at Swift. “That means collaborating to overcome industry challenges and ultimately deliver a better experience to customers. Whether your institution with an idea to transform the future of screening, eliminate payment friction or better connect corporates, get in touch.”

Get in touch

Loading...