Building AI applications in Rust

Rust is rapidly becoming a serious platform for building AI-powered systems, but what does shipping real LLM applications in Rust actually look like day to day? How do agents, tooling, and workflows come together in practice? In this livestream, Rust Developer Advocate Orhun Parmaksız is joined by Stephen Korzeniewski, Lead Maintainer of Rig at 0xPlaygrounds, to explore how Rust and AI meet in real-world development. They’ll walk through how Rig fits into the Rust + AI ecosystem and how Stephen uses it to build modular, scalable and ergonomic LLM-powered applications in Rust. If you’re interested in using Rust for AI, building agents, or understanding what modern Rust+LLM workflows look like in practice, this session offers a deep dive from the person leading one of the most active Rust AI frameworks. #Rust #AI #LLM #Agents #OpenSource #developertools 00:00 Introduction & livestream series context 02:06 Stephen's background 04:59 The early AI ecosystem 10:27 Ratatui, Bevy & VTuber setup 14:55 Why Rust works well for AI? 23:19 What is Rig? 25:07 Becoming Rig's lead maintainer 26:51 Minimal terminal coding agent demo 42:01 Rust's type system, traits and abstraction 45:41 Multi-agent systems in Rig 47:33 Vector databases and RAG 52:54 Testing nondeterministic LLM applications 56:37 Rig Candle & local inference 58:37 Rig in production 01:02:35 Downsides of using Rust for AI applications 01:04:54 Stephen's AI-first programming workflow 01:07:00 AI slop & software quality 01:09:31 Context switching and managing multiple coding agents 01:12:59 Reviewing AI-generated code 01:14:36 The satisfaction of programming 01:17:21 Correct code and technical knowledge 01:21:24 Why developers oppose AI? 01:27:26 AI dependency, cost, local models and token limits 01:30:01 How coding-agent tooling may evolve 01:32:27 The near future of Rust and AI 01:36:30 AI contributions to open source 01:40:54 Avoiding endless AI code-review 01:43:38 Closing remarks