TD Bank: Enable data & machine learning scientists with Posit & Databricks

In the realm of financial institutions, the utilization of data, particularly in the realms of Artificial Intelligence (AI) and Machine Learning (ML), is crucial for solving a myriad of challenges that these institutions face. However, within these institutions, which handle extremely sensitive data, the top concerns from leaderships often revolve around security and the intricate communication dynamics across diverse teams, multiple networks, and intricate leadership structures. Crafting products that are not only functional for end-users but also guarantee the safeguarding of client data presents a formidable task for these teams.

At the forefront of navigating these challenges is Mehran Moghtadai, the head of AI/ML Accelerator & Enablement at TD Bank. Leading a team of engineers, Mehran focuses on delivering high-value use cases tailored for businesses dealing with vital and sensitive information.

Starting at [23:15], Mehan shares a few tools he uses within his team at TD Bank:

• Spark
• Hive
• Posit Workbench
• Jupyter Labs
• Posit Connect
• Azure

Mehran shares at [30:30] a few tips for implementing open source tools in a high risk environment where leadership may be hesitant:

• Develop partnerships with enterprise protection teams to scan libraries
• Everything is vetted by a model validation team.
• Build trust in businesses and leadership.

These controls are put in place in order to use open source tools in a regulated environment.

Another recommendation from Mehran at [43:05] is: “the key here is to be able to abstract your use case into a pattern… that you can explain to the partners and say, this is what I need to do.” This includes understanding that partners and leadership will ask questions about data security and being ready to establish the infrastructure that helps do that.

The intersection of data-driven technologies and financial institutions necessitates a delicate balance between innovation and security. Mehran Moghtadai’s experiences shed light on the importance of trust-building, effective communication, and a nuanced understanding of security protocols to successfully navigate the challenges inherent in developing and maintaining AI/ML projects within this dynamic landscape.