Your Privacy

This site uses cookies to enhance your browsing experience and deliver personalized content. By continuing to use this site, you consent to our use of cookies.
COOKIE POLICY

Skip to main content
Back to insights

Bridging Clouds: Building a Multi-Model LLM Chatbot with Azure OpenAI and Amazon Bedrock

March 19, 2025

Science Museum of Virginia – Dewey Gottwald Center
Watch Recording

Join us at the 2025 rvatech/ Data + AI Summit to hear our Chief Technology Officer, Josh Bartels, present on building multi-modal LLM chatbots. 

In today’s rapidly evolving AI landscape, the ability to harness multiple cloud providers and language models is becoming increasingly crucial. This session will explore the journey of the creation of a Multi-Model LLM Chatbot that utilizes both Azure OpenAI and Amazon Bedrock.

Key highlights of this talk include:

  1. Architecture Overview: Explore the intricate architecture behind a multi-cloud, multi-model chatbot. We’ll discuss how to design a system that can seamlessly switch between different LLMs based on specific needs.
  2. Integration Showcase: Watch as we demonstrate live integration with Azure OpenAI’s GPT models and Amazon Bedrock’s array of foundation models. Learn how to leverage the strengths of each platform to create a more robust and versatile chatbot. We’ll cover authentication, API calls, and how to handle the unique features of each service.
  3. Live Chatbot Demo & Code Review: Experience the Multi-Model LLM Chatbot in action through a live demonstration. We’ll interact with the chatbot, showcasing its ability to seamlessly switch between Azure OpenAI and Amazon Bedrock models. Following the demo, we’ll dive into a code review of crucial components, examining the logic behind model selection, request handling, and response processing. This segment will provide you with practical insights into implementing a multi-model system.
  4. Best Practices & Pitfalls: Gain insights into best practices for multi-model implementations, common pitfalls to avoid, and strategies for maintaining consistency across different LLMs. We’ll discuss prompt engineering techniques, error handling, and how to ensure a coherent user experience despite using multiple models.

About Josh Bartels

With over 20 years at the forefront of technology innovation, I've dedicated my career to delivering strategic solutions that drive business growth. As the Chief Technology Officer of UDig, I lead our technology vision, architecting solutions that transform how organizations leverage technology to generate impact.

Digging In

  • Artificial Intelligence

    Four Things That Stuck With Me From NAW SHIFT

    I spent last week in Denver at NAW SHIFT, my first one as an affiliate member. Across the sessions I sat in on, four themes kept coming up in different ways. AI was one of them, but honestly, it was not the most interesting one. The better conversations were about what has to be right underneath the technology before any of it […]

  • Artificial Intelligence

    UDig AI Week: We Cleared the Calendar & Went All In

    For one week, UDig’s entire delivery team — engineers, designers, and product owners — put client work down and focused on one thing: rebuilding how we work with AI. We didn’t plan AI Week because we had a perfect strategy. We planned it because we looked at where software delivery is heading — the disappearing middle, the compression of […]

  • Artificial Intelligence

    Coding at 5x: How AI Boosted Our Team’s Productivity

    When I look back at my career in technology one day, I believe that 2025 will be the year I reflect upon as when everything changed. There was software engineering pre-2025 and software engineering post-2025. Late 2024 is when the AI-Powered IDE’s, specifically Windsurf and Cursor, came to market. Quickly followed by agentic coding tools […]

  • Artificial Intelligence

    Automating Discovery: Turning Requirements into Jira Stories with AI

    When UDig was asked to explore ways to accelerate delivery, the brief was intentionally open-ended, inviting the team to rethink existing processes and challenge assumptions. One area quickly emerged as a clear opportunity: discovery. While essential, discovery can slow momentum when large volumes of requirements must be manually translated into user stories. Like most projects, […]

  • Artificial Intelligence

    Generative BI: Building a Natural-Language Analytics Engine

    Our recent exploration into generative analytics uncovered exciting possibilities for the future of business intelligence. We set out with a broad goal: to democratize analytics insights and eliminate bottlenecks by giving users a personal data analyst. The result was GenBI, an internal proof of concept demonstrating how large language models can sit on top of structured datasets, translate natural language into SQL, and generate accurate charts in […]

  • Artificial Intelligence

    Agentic Commerce: Four Paths Retailers Can Take Right Now

    With over 40% of shoppers saying AI is now their primary source of insight, today’s agentic commerce tools create unprecedented visibility into consumer purchase intent and decision-making patterns. Today’s AI agents excel at surfacing clear product data and creating frictionless shopping experiences. Smart retailers already recognize agentic commerce as a differentiation opportunity, and some major […]