Rustem Feyzkhanov · Writing
Re:Invent 2023 attendee guide: Generative AI
Attendee guide: Generative AI
Explore the top 17 AWS re:Invent 2023 breakout sessions to stay on top of everything related to generative models.
Machine Learning Hero
Machine Learning Engineer at Instrumental
Generative AI is starting to become an essential part of many companies. There are challenges with deploying generative AI solutions to productions since architectures are different from the traditional MLOps. Here are the top 17 re:Invent sessions to help you stay on top of everything related to generative models.
This year, re:Invent will have both in-person and virtual experiences. The virtual experience is free and you can register for virtual attendance on the AWS Events Portal for access to the keynote and innovation talk livestreams. Most breakout session recordings are available on YouTube after they’re presented.
Recommendations
LLM intro and tutorial sessions
These sessions will bring you up to speed in the large language model (LLM) world and will show you how to utilize AWS infrastructure for small LLM projects.
COM203 | Unlocking your full potential with the power of generative AI on AWS
Generative AI allows machines to produce diverse content, and AWS offers tools to harness its potential. This session delves into how Amazon SageMaker and Amazon Bedrock help developers integrate generative AI into their applications, fostering unprecedented creativity and business transformation.
COM201 | Building a cloud-backed generative AI game in 60 minutes
Discover the potential of generative AI in creating unique user experiences. This session showcases a demo of constructing a game with Amazon Bedrock, Amazon API Gateway, and AWS AppSync where players guess movie or book titles from emoji plot summaries powered by Amazon Titan.
COM202 | Building an AI comic video generator with Amazon Bedrock
Discover the journey of AWS Machine Learning Hero Agustinus Nalwan in creating Owly, a unique AI-driven bedtime storyteller. This session delves into the use of Amazon Bedrock for script generation, fine-tuning with Amazon SageMaker JumpStart, and advanced techniques in prompt engineering and image blending with the Stable Diffusion model.
RECOMMENDATIONS
LLM-related architectures
These sessions address the technical intricacies and architectural considerations of integrating LLMs and generative AI in various environments.
SMB301 | Building inventive customer experiences with data, AI, and serverless
Modern businesses thrive by tailoring customer experiences with data, machine learning, and generative AI, leading to increased revenue and loyalty. This session dives into building these experiences using data, AI, and AWS serverless technology, highlighting design strategies, success metrics, and AWS solutions to guide you.
SAS306 | SaaS meets AI/ML & generative AI: Multi-tenant patterns & strategies
Navigating AI/ML integration in a SaaS setting involves understanding multi-tenant challenges and effective mapping of tenants to models. This session delves into the intricate relationship between multi-tenancy and AI/ML, offering insights on design, performance, and strategies for seamless AI/ML SaaS integration.
CMP201 | Advancing generative AI innovation with Amazon EC2 accelerated compute
Discover how generative AI can revolutionize applications, customer experiences, and productivity using the Amazon EC2 optimized instances. This session delves into the capabilities and benchmarks of these instances and shares insights from AWS customers on deploying large-scale text and image generation models.
DAT407 | Best practices for querying vector data for gen AI apps in PostgreSQL
Discover how to harness PostgreSQL’s pgvector extension to optimize vector data storage and querying for AI/ML applications. This session will guide you on storing Amazon SageMaker data in Amazon Aurora PostgreSQL-Compatible Edition and on fine-tuning SQL queries for peak performance with AI/ML data and vector-specific tools.
CON312 | Navigating the future of AI: Deploying generative models on Amazon EKS
Discover the transformative impact of generative AI on business-customer interactions and the role of open source tools on Kubernetes. This session delves into tools such as Ray, JupyterHub, Argo Workflows, and Karpenter, highlighting how they enhance generative AI application deployment on Amazon EKS with real-world use cases.
STG212 | Build your generative AI data strategy from your data lake storage
Discover the latest trends in AI, machine learning, and how they’re shaping new storage opportunities in various industries. This session will guide you through AWS storage solutions designed for machine learning, focusing on high-performance, scalability, and seamless integration with a variety of AI/ML services.
Recommendations
LLM use cases
An important part of designing great architecture for LLMs is understanding success stories for implementing them. That helps with making sure you design a solution that is scalable, cost-efficient, and delivers value.
PRO301 | Next-gen trucking: How Iveco uses AWS to harness generative AI
Iveco is using AWS-based generative AI to enhance the connection between drivers and vehicles with privacy-compliant data. This session delves into how these innovations improve the Iveco driver experience and the pivotal part AWS plays in their responsible implementation.
GBL201-DE | Generative AI: From individual use to enterprise application [German]
Generative AI in 2023 brought transformative changes with potential annual benefits of $4.4 trillion in various sectors. This session, delivered in German, delves into its use cases, addresses challenges, offers solutions using AWS tools, and highlights country-specific considerations for application development.
GBL209-KO | Lessons learned using generative AI in banking and shopping services [Korean]
Korean companies utilize generative AI for innovation in banking and shopping, including AI-driven call centers and advanced product recommendation systems. This session, delivered in Korean, delves into KB Financial Group’s AI-powered call routing and EMART’s use of Amazon SageMaker and Amazon Rekognition for personalized online shopping experiences.
HLC204 | Improving patient outcomes using generative AI in healthcare
Generative AI is revolutionizing healthcare by enhancing care quality and diagnosis speed at the University of California San Diego Health. This session delves into how clinicians utilize Amazon SageMaker and generative AI on Amazon Bedrock to analyze interventions and optimize patient outcomes by aligning comorbidities with patient demographics.
IMP202 | Innovating for scientific discovery and health equity with AI/ML
Discover the potential of AI-powered applications and open data models in driving transformative impacts in healthcare and scientific discovery. This session delves into an organization’s use of generative AI to aid cancer patients in underserved regions and explores the evolution of ethical and efficient AI models for science and literacy.
LFS202 | Accelerating life sciences innovation with generative AI on AWS
Life sciences companies are using generative AI, like Amazon Bedrock, to enhance innovation and streamline operations. This session will focus on how Gilead employs this technology to optimize clinical trials and delve into its conversational UI, content summarization, and semantic search features.
Recommendations
Generative AI DevTools
Generative AI–powered DevTools transform normal development workflows and significantly increase productivity. In these sessions, you’ll discover the latest AWS tools and technologies that use generative AI to improve the software development lifecycle.
DOP201 | Best practices for Amazon CodeWhisperer, an AI coding companion
Generative AI enhances developer productivity by creating content, from conversations to code suggestions. This session delves into best practices for using Amazon CodeWhisperer, a generative AI tool that offers real-time code assistance.
BOA303 | Manage your data more efficiently with AI using Amazon CodeWhisperer
Learn about optimizing data management and the machine learning lifecycle using Amazon SageMaker and Amazon CodeWhisperer. This session covers how SageMaker aids in dataset preparation, model building, training, and deployment, while the CodeWhisperer extension streamlines tasks like data ingestion, query formulation, and visualization.
DOP207 | Build and run it: Streamline DevOps with machine learning on AWS
Developers often face challenges with manual code reviews, multi-language proficiency, online code snippet searches, and application observability. This session teaches you how to use machine learning tools like Amazon CodeWhisperer and Amazon DevOps Guru to enhance application availability and improve software development speed and reliability.