Witad Awards 2020: Data Science Professional of the Year—Anita Patel, Bank of America

Anita Patel has never been afraid of uncharted territory. As a teenager at school in London, there were few guides to help her pursue a career in technology, but now she is a lead data scientist in sales, research and capital markets at Bank of America Securities. And then, having embarked on this career, she pioneered the use of predictive models within that business.

Patel started her career at the bank as a software developer in a different business, where she began to work with predictive models. “I was introduced to the bank’s capital markets business and they were not so used to using technology like machine learning and building out predictive models. So we had a lot of fun exploring this area—it was completely uncharted,” she says.

The model Patel built, the Predictive Intelligence Analytics Machine (PRIAM), won the 2018 American Financial Technology Award for the most cutting-edge initiative. PRIAM is an artificial intelligence system that identifies within seconds the investors most likely to participate in an IPO or block deal, helping to maximize demand for equity capital market transactions. PRIAM has run predictions on equity capital markets transactions totaling more than $1 billion in aggregate deal volume to date, with 80% accuracy on the top 30 investors in a deal.

With PRIAM, BofA Securities was able to improve the process of finding investors, an undertaking analysts had been doing manually and based on anecdotal analysis for years.

“It would take analysts several weeks. And we could gather all of this data, and apply a model that would generate a list for them. They could still review it before sharing it with clients to ensure they agree with the predictions, but it would simplify the process a lot for them,” Patel says. “Our analysis enabled our bankers to find incremental, ‘high signal’ investors who may not have been initially identified in their analysis as being interested in a particular deal.”

Apart from her work on PRIAM, Patel has also developed time-series forecasting models, such as sales trading shifts, for areas of the bank outside of capital markets. She wants to help women land STEM jobs, so she works with organizations like Stemettes and Code First: Girls, and also participates in recruitment events at her alma mater, Imperial College London. Within Bank of America, she has organized and hosted a data science workshop in collaboration with the Women’s Developers Network, which aimed to teach developers and non-developers how to create predictive models.