Mastering Model Inference Strategies on AWS

Oussama Kandakji | Sr. AI & ML GTM Solutions Architect at AWS

This session explores AWS model inference strategies, covering batch processing, multi-model hosting, and serverless options. Learn to optimize performance, reduce costs, and meet diverse inference needs. Topics include batch inference for large datasets, multi-model hosting on single infrastructure, serverless and event-driven services for scalable, on-demand inference.

Coffee Break

Own, Control, and Optimize your AI Models at Scale

Kevin FRANCOIS-BOUAOU & Oussama KANDAKJI | Manager AI/Generative AI at Deloitte & Sr. AI & ML GTM Solutions Architect at AWS

Covering strategies to effectively manage and optimize foundation models at scale. Topics: model quantization, formats, orchestration for deployment/training, capacity usage, MLflow. Implementing scalable solutions for foundation models in production. Practical guidance to unlock their potential.

Lunch

The art of ML automation (MLOps) using MLflow on Amazon Sagemaker

Christian KAMWANGALA & Ioan CATANA | Big Data Cloud Engineer & AI/ML Specialist Solutions Architect at AWS

In this session, we'll explore how to leverage MLflow, an open-source ML lifecycle platform, on Amazon SageMaker to automate key ML workflow stages, including experiment tracking, model management, and deployment. Through hands-on examples, you'll learn how to construct a robust, scalable MLOps pipeline that improves the efficiency, reproducibility, and governance of your ML projects.

Coffee Break

Serve 100s of fine tuned llms for the price of 1 with Amazon Sagemaker

Hossam Basudan | Solutions Architect at AWS

This session explores serving multiple fine-tuned language models cost-effectively using Low-Rank Adaptation (LoRA) on Amazon SageMaker. Learn how LoRA enables efficient fine-tuning with significant savings. Discover a multi-adapter serving solution for deploying diverse fine-tuned models on a single base model, minimizing performance and cost impact.