Coffee with Developers - Maria Apazoglou

How do you build a secure AI platform for 9,000 employees? Learn how Thomson Reuters democratized generative AI while building user trust and managing expectations.

#1about 2 minutes

Tracing a career path in data and AI platforms

Maria shares her journey from creating early models to leading AI and data platforms at major financial and information services companies.

#2about 2 minutes

Shifting the AI platform focus to all employees

The rise of generative AI prompted a strategic shift from serving only data scientists to enabling all employees to create their own AI solutions.

#3about 4 minutes

Driving AI adoption with security and self-service

A self-serve UI, a security-first multi-account architecture, and extensive internal training were key to achieving widespread user adoption.

#4about 2 minutes

Managing AI hallucinations and cost transparency for users

A multi-model comparison feature and transparent cost reporting help users understand AI limitations and make informed decisions.

#5about 5 minutes

Improving efficiency with reusable custom AI solutions

Users can create and save custom prompt "chains" to automate repetitive tasks, turning conversational AI into a reusable productivity tool.

#6about 6 minutes

Grounding AI responses with data and citations

To ensure factual accuracy, the platform encourages users to provide their own data and uses prompt engineering and citations to ground model outputs.

#7about 3 minutes

The value of cognitive diversity in building AI teams

Diverse teams with varied backgrounds, skills, and perspectives are crucial for building robust and user-friendly AI products.

#8about 5 minutes

How AI creates new career paths for subject matter experts

AI tools enable subject matter experts to transition into roles like prompt engineering, while core engineering challenges shift to scale and performance.

#9about 4 minutes