In anticipation of our first Ampere® Developer Summit on Thursday, September 26th, we're spotlighting our guest speakers and diving into their topics and insights. If you haven't signed up yet, there's still time!

Register Now

Last week, we introduced the first session in the upcoming Ampere Developer Summit: “Why Ampere CPUs Have an Impact for Application Developers.” This week, we’re taking a closer look at another important question: “How Can You Efficiently and Affordably Run Your AI and ML Models in Production on Ampere CPUs?”

The Challenge of Moving AI from Hype to Production
AI is a buzzword in both the tech world and beyond, with nearly every company rebranding itself as an AI company. However, AI appears to have passed through Gartner's Peak of Inflated Expectations and is now entering the Trough of Disillusionment. This transition happens because moving from AI training to production is notoriously difficult—nearly 80% of models that exit training never make it into production. Additionally, even when models do make it to production, they are often expensive to run, making them not cost-effective.

To help AI transition from the Trough of Disillusionment to the Slope of Enlightenment, it's essential to reduce the total cost of ownership (TCO) by simplifying, streamlining, and lowering the costs of creating, training, and deploying models in production (AI inference). But how can this be achieved?

The Solution: Discover Ampere and Oracle's Approach
Our experts will guide you through how to effectively and efficiently build AI models, making deployment to production faster and easier. Once in production, you can run optimized large language models (LLMs) without relying on expensive, energy-inefficient, and often scarce GPUs. Instead, you’ll use off-the-shelf, energy-efficient, AI-optimized Cloud Native Processors like Ampere’s A1 and A2 instances on OCI, allowing you to save both time and money while advancing towards the Plateau of Productivity.

Event Speaker Lineup

• Victor Jakubiuk, Head of AI at Ampere Computing, will walk you through the evolution of large language models (LLMs), current AI trends, and future projections.
• Amar Gowda, Senior Principal Product Manager at Oracle, will focus on how Oracle Cloud Infrastructure (OCI) uses Ampere CPUs to deliver cost-effective AI and ML solutions. Amar will demonstrate how Oracle’s AI initiatives, powered by Ampere CPUs, offer developers a scalable, energy-efficient, and budget-friendly approach to deploying AI models into production without the need for expensive GPUs.
• Steve Notley, Director of Field Engineering at Wallaroo.AI, will discuss how Wallaroo's MLOps platform helps companies seamlessly move AI models from research to production. Steve will share how Wallaroo has seen better performance running on Ampere-based OCI A1 instances, particularly when using Ampere Optimized AI Frameworks (AIO) and Ampere Optimized llama.cpp. This session will highlight the ongoing challenge of getting models deployed and how Wallaroo’s platform simplifies that journey.

Speaker Profiles

• Victor Jakubiuk, Head of AI at Ampere Computing, co-founded OnSpecta, a software AI acceleration company, and served as its leader for 8 years until it was acquired by Ampere in 2021. He now leads Ampere's AI group.
• Amar Gowda, Senior Principal Product Manager at Oracle, focuses on AI/ML incubations. His team drives innovative generative AI and AI frameworks initiatives for Oracle Cloud Infrastructure (OCI).
• Steve Notley, Director of Field Engineering at Wallaroo.AI, has been working in AI/ML for over 8 years. He earned his master’s degree from Rensselaer Polytechnic Institute with a focus on machine learning and currently leads Wallaroo’s Field Engineering team.

Don’t miss the opportunity to learn how Ampere and Oracle are reshaping AI production. Join us at the Ampere Developer Summit to gain valuable insights from industry experts on making AI more efficient, affordable, and scalable. Sign up today and be part of the conversation!