BQuant “Code Crunch” Hackathon in London Showcases Cutting-Edge Quant Finance Innovation

Teams from T. Rowe Price, Atlantic House, J.P. Morgan and Bloomberg use BQuant Enterprise to push the boundaries of what's possible in quant finance

LONDON - Bloomberg recently hosted a BQuant Code Crunch Hackathon in its EMEA Headquarters in London, bringing together teams from leading financial institutions to develop and present pioneering solutions to real-world financial challenges. Dedicated to innovation, collaboration, and technical excellence, the event featured some of the brightest minds in quantitative finance and machine learning (ML).

Powered by Bloomberg's BQuant Enterprise platform, "Code Crunch" was designed to foster experimentation, encourage relationship-building, and enable knowledge sharing across the quant finance community. It concluded a rigorous four-month program featuring extensive collaboration, coding sessions, workshops, and deep technical engagement by the participating teams.

"The BQuant Code Crunch Hackathon reaffirmed the power of collaboration in driving innovation," said Javier Mitra Valdes, Head of Product for the BQuant Platform at Bloomberg. "With machine learning and AI continuing to reshape the financial landscape, events like these are vital for fostering the next generation of quant solutions."

Spotlight on Innovation: Team Presentations
The event showcased advanced quantitative finance and machine learning applications the teams built using BQuant Enterprise to transform complex financial data into intuitive visual tools, including sentiment analysis, economic forecasting, and news-driven analytics. Each team had 10 minutes to present their projects, followed by a Q&A session. These included:

• T. Rowe Price: Julien Fabre, Team Leader, Fixed Income Systematic Research, presented a sentiment analysis model that uses BERT and generative AI to track and analyze Bloomberg News sentiment to forecast market dynamics related to tariffs. This project demonstrated how event-time analysis can better capture market dynamics than traditional calendar-based approaches.
• Atlantic House: Mark Greenwood, Deputy Chief Investment Officer, showcased a sophisticated analytical tool that mines economist estimates to improve inflation forecasts and identify potential alpha through advanced machine learning techniques. By ranking economists based on historical accuracy and bias, the team achieved a significant improvement in mean absolute error - translating to meaningful alpha in macro strategies.
• J.P. Morgan: Daniel Thiel, Executive Director, Quantitative Research, Equities, introduced a profit warning summarization tool that utilizes PySpark and Bloomberg's fine-tuned T5 NLP model to process and summarize Bloomberg News headlines. The application automated a previously manual process, offering real-time insights into market-moving news with a user feedback loop for continuous model refinement.
• Internal Bloomberg team: Market Specialist Anish Popat and Desktop Build Group analyst Ricard Radomski demonstrated an ML framework that uses neural networks, random forests, and scalable compute infrastructure to construct smarter equity signals. Their modular API enables rapid deployment and backtesting, improving the efficiency and effectiveness of quantitative analysis workflows.

Award Winners
The event concluded with an awards ceremony recognizing outstanding contributions:

• Visual Insighter Award - J.P. Morgan: For creating an intuitive and interactive UI that made complex data accessible, actionable, and deeply understandable at a glace
• Solution Builder Award - Bloomberg: For constructing a robust workflow that brought together data, models, and interactivity into a cohesive and impactful application
• Data Innovator Award - T. Rowe Price: For their innovative and skillful use of Bloomberg News data and generative AI in sentiment analysis
• Model Master Award - Atlantic House: For developing a deep and powerful model that demonstrates expert-level application of machine learning or NLP with a balance of performance, interpretability, and robustness
• Alpha Innovator Award - Ampega Asset Management (in absentia): For their novel use of available data and capabilities to build an NLP-powered bond optimization tool helps with alpha generation

The award recipients were decided upon by a panel of judges from Bloomberg's Office of the CTO: Bruno Dupire, Head of Quant Research, Gary Kazantsev, Head of Quant Technology Strategy, and Arun Verma, Head of the Quant Research Solutions Team.

Dupire commented: "As one of the initiators of the BQuant initiative, I'm excited to see how this solution is gaining a lot of traction and the level of engagement and involvement by the community. Every project brought unique value to the table, and it's inspiring to see how far we've come."

The Code Crunch Hackathon series underscores Bloomberg's commitment to fostering innovation through collaboration, reflecting its vision for BQuant Enterprise as a foundational tool for scalable, flexible, and impactful quantitative research. Additional BQuant "Code Crunch" Hackathons took place in the second half of 2025 in Beijing, Shanghai, Tokyo and Singapore.

Read a detailed report about the BQuant projects built during the Code Crunch Hackathon in London at https://www.bloomberg.com/professional/insights/financial-services/code-crunch-honors-data-innovators/.