Kesri Tavadia, CIO at Bombay Stock Exchange has over 35 years of vast industry experience in IT – specifically software development, infrastructure services (networks, OS and databases) and cybersecurity for financial services and capital market firms. In his nine-year stint at BSE, Tavadia has played a leading role to transform BSE into the world’s fastest exchange. He is a strong proponent of open source technologies and has walked the talk by successfully implementing one of the largest data warehouses on open source platform. Providing technical leadership and architectural vision is core to his heart; under his leadership at BSE, numerous mission-critical projects have been implemented – for instance, an automated complete disaster-recovery procedure.

Stock Exchanges are most impacted with real time news of the listed companies especially with rumors that often sky rocket or plummet the share prices of companies. BSE Limited implemented artificial intelligence & machine learning in its Surveillance & Investigation function to identify rumors about BSE listed companies in near-real time that analyzes live streaming inputs from various TV channels. The project involved automated speech recognition (ASR) with the help of ML and AI based system by applying deep learning algorithm for natural language processing end-to-end speech recognition or ASR that converts a speech signal to a sequence of words. The system built using open source framework and big data technologies has no licensing cost, which reduces the cost burden on BSE to a large extent. The project was implemented for the surveillance department, but with the increase on the quantum of processing information, the department was able to monitor impact on prices. Employee productivity saw a spike as the team involved were equipped with better and more information. There was overall cost savings in terms of resources deployed for manual tracking. The social media analytics system was able to predict rumors about BSE listed companies with a 99 percent success ratio. Tavadia explains that the accuracy of AI based systems using DL module will further increase as the system is continuously trained with large volumes of data over time. The use of AI and ML in surveillance and investigation function at BSE helped maintain the integrity of the stock market, in addition to increasing the brand value of BSE.