Cloud Masters Episode #133
GenAI at production scale: Why GenAI POCs fail and how AWS thinks about production readiness
Covering why GenAI projects fail during POC-to-production transitions and AWS Bedrock frameworks for successful enterprise GenAI deployment.
Cloud Masters Episode #133

With DoiT Spot Scaling, automate your AWS Spot Instances to save up to 90% on compute spend without compromising reliability.

Cloud Masters
Cloud Masters
GenAI at production scale: Why GenAI POCs fail and how AWS thinks about production readiness
Loading
/
Cloud Masters
Cloud Masters
GenAI at production scale: Why GenAI POCs fail and how AWS thinks about production readiness
Loading
/

Episode notes

Key Moments

00:00: Introduction
01:38: The evolution of GenAI adoption
04:45: Why successful GenAI PoCs fail at production scale
08:25: FinTech and insurance GenAI migration stories
12:36: Four patterns of GenAI workload migration
18:38: Amazon Bedrock migration framework
21:50: LLM performance evaluation strategies
27:10: From model assessment to GenAI deployment
29:35: Enterprise AI compliance challenges
34:43: AI chatbot risks and Bedrock guardrails
40:35: What makes GenAI migrations successful
46:40: LLM ecosystem costs beyond model usage

About the guests

Ninad Joshi
Ninad Joshi is a GenAI Partner Solutions Architect at Amazon Web Services, where he spearheads the adoption of AWS’s generative AI services across strategic partnerships throughout EMEA. Ninad has spent over three years at AWS leading cutting-edge AI/ML initiatives, including work on Amazon Bedrock. He’s recognized for his thought leadership in generative AI, having published innovative architecture patterns that help organizations harness the power of AI technologies.
Rajan Bhave
Rajan Bhave, a Data & AI Specialist at DoiT International, is based in the Stuttgart region of Germany. He is deeply passionate about AI, data, and cloud technologies, consistently seeking to learn and share insights in this dynamic field. Rajan values connecting with others, exchanging ideas, and fostering continuous learning. In his free time, he enjoys cooking, exploring trails, hiking, swimming, and playing table tennis.
Ninad Joshi is a GenAI Partner Solutions Architect at Amazon Web Services, where he spearheads the adoption of AWS’s generative AI services across strategic partnerships throughout EMEA. Ninad has spent over three years at AWS leading cutting-edge AI/ML initiatives, including work on Amazon Bedrock. He’s recognized for his thought leadership in generative AI, having published innovative architecture patterns that help organizations harness the power of AI technologies.
Rajan Bhave, a Data & AI Specialist at DoiT International, is based in the Stuttgart region of Germany. He is deeply passionate about AI, data, and cloud technologies, consistently seeking to learn and share insights in this dynamic field. Rajan values connecting with others, exchanging ideas, and fostering continuous learning. In his free time, he enjoys cooking, exploring trails, hiking, swimming, and playing table tennis.

Related content

Data prerequisites for building GenAI applications
In this clip from our webinar on GenAI applications in Financial Services, AI/ML expert Rupal Bhatt discusses the essential prerequisites for working with data in GenAI applications.
Watch now
The cost impact of Large Language Models (LLMs) in production
We cover the ever-growing importance of Large Language Models (LLMs) in applications, how LLM costs can easily compound once in production, and breaking down the costs associated with using LLMs.
Watch now
Anatomy of an LLM
Large language models (LLMs) like Claude, Cohere, and Llama2 have exploded in popularity recently. But what exactly are they and how can you leverage LLMs to build impactful AI applications? This article is an in-depth look at LLMs and how to use them effectively on AWS.
Read more

Schedule a call with our team

You will receive a calendar invite to the email address provided below for a 15-minute call with one of our team members to discuss your needs.

You will be presented with date and time options on the next step