Skip to main content
AI Engineer World's Fair (2024)

Keynotes: Part 2


Head back to all of our AI Engineer World's Fair recaps

Unlocking Developer Productivity across CPU and GPU with MAX

Chris Lattner @clattner_llvm / Modular
Watch it on YouTube | AI.Engineer Talk Details

This talk introduces Modular's Max framework and Mojo programming language, which aim to simplify and accelerate AI development and deployment. The speaker discusses how these tools address challenges in AI engineering by offering easier GPU programming, improved performance, and a unified stack that combines the benefits of Python with high-performance computing capabilities.

From Software Developer to AI Engineer

Antje Barth @anbarth / AWS
Watch it on YouTube | AI.Engineer Talk Details

This talk by an AWS representative outlines five practical steps for software developers to become AI engineers, focusing on understanding AI fundamentals, using AI developer tools, prototyping with AI, integrating AI into applications, and staying up-to-date with the rapidly evolving field. The speaker showcases AWS tools like Amazon Q Developer and Amazon Bedrock, and demonstrates how to build AI-powered applications and agents, including a Minecraft-playing AI bot.

What's new from Anthropic and what's next

Alex Albert @alexalbert__ / Anthropic
Watch it on YouTube | AI.Engineer Talk Details

This talk by Alex Albert from Anthropic discusses the current state of AI integration in products, comparing it to historical technological transitions, and introduces Anthropic's latest advancements including the Claude 3.5 Sonnet model and new features like artifacts and projects. He emphasizes the need to redesign products from the ground up with AI capabilities in mind, rather than simply adding AI features to existing structures, and previews upcoming developments in Anthropic's AI offerings.

LangChain Launch: Infrastructure for building reliable agents

Harrison Chase @hwchase17 / LangChain
Watch it on YouTube | AI.Engineer Talk Details

This talk by a representative from Langchain introduces LangGraph, a stable version of their framework for building custom AI agent architectures, and announces LangGraph Cloud, a new service that combines LangGraph's flexibility with production-ready infrastructure for deploying AI agents. The speaker also previews LangGraph Studio, a tool for building, debugging, and sharing AI agents, emphasizing the importance of customizable cognitive architectures and human-in-the-loop features for production-ready AI agent systems.

From Text to Vision to Voice: Exploring Multimodality with OpenAI

Romain Huet @romainhuet / OpenAI
Watch it on YouTube | AI.Engineer Talk Details

This talk by Romain from OpenAI showcases the company's latest advancements in AI, particularly focusing on GPT-4 with Vision and multimodal capabilities, including live demonstrations of voice interactions, image analysis, and coding assistance. The speaker also outlines OpenAI's future focus areas, including improving textual intelligence, developing faster and cheaper models, enabling model customization, and advancing AI agents, while emphasizing the company's commitment to supporting developers in building AI-native applications.

⭐ What We Learned From A Year of Building With LLMs

Eugene Yan @eugeneyan / Amazon
Hamel Husain @HamelHusain / Parlance Labs
Jason Liu @jxnlco / Independent & Instructor
Dr. Bryan Bischof @bebischof / HEX
Charles Frye @charles_irl / Modal
Shreya Shankar @sh_reya / UCB EECS & EPIC Lab, UC Berkeley
Watch it on YouTube | AI.Engineer Talk Details

This talk, presented by six co-authors on building with LLMs, covers strategic, operational, and tactical considerations for developing AI applications. The speakers discuss the importance of focusing on product expertise rather than model development, the need for continuous improvement cycles, challenges in hiring AI engineers, and practical advice on implementing evaluations, data analysis, and guardrails when working with LLMs.

Read our Deep Dive on this talk as well.

Copilots Everywhere

Thomas Dohmke @ashtom / GitHub
Swyx @swyx / Latent.Space / Smol
Watch it on YouTube | AI.Engineer Talk Details

This talk features Thomas Dohmke, CEO of GitHub, discussing the development and impact of GitHub Copilot, as well as the company's vision for AI-assisted software development. He emphasizes how AI tools like Copilot and Workspace aim to enhance developer productivity, democratize access to coding, and bring more enjoyment to software development while maintaining a human-centric approach.