AWS Innovation with Epsilon Digital Marketing

Topic: Scalable Serverless Taxonomy Solutions

About the Episode

On this episode of Innovation Ambassadors, we're showcasing the journey of Epsilon Digital Marketing - a global leader in outcome-based marketing, helping the world's top brands personalize experiences for millions of people.

Get a behind-the-scenes look at developing a distributed system to rapidly categorize massive product datasets across clients with services like Amazon EMR, and Amazon Bedrock. Hear about their iterative process enhancing the AI models, improving throughput, upskilling on modern architectures, and fusing marketing domain expertise with AWS's technical capabilities - all while delivering highly personalized ad experiences at epic scale with privacy-by-design.

Highlights

00:00 - Introduction
01:08 - Epsilon Digital Marketing's business of serving personalized ads at massive scale
03:48 - Epsilon's complex legacy technology landscape
06:56 - How Epsilon partnered with AWS Prototyping team
09:56 - The three core taxonomy processes modernized

12:09 - The two-step prototyping approach: understanding legacy first, then applying Gen AI
15:02 - Benefits of generative AI reducing manual taxonomy work
18:45 - The importance of "privacy by design" principles at Epsilon
22:05 - Overcoming obstacles and learnings during the prototype project
29:28 - AWS prototyping is hiring!

The barriers to entry were less than I anticipated. And so, if you can find the right business use case within your organization, that you think Gen AI would make a difference for and help you evolve your current capabilities. I'd say, just jump in the pool, right?"

John Conley encourages others to explore generative AI for suitable use cases, finding it more accessible than expected.