Rust at OpenAI: Performance, Safety and Scale @JetBrainsTV
Rust at OpenAI: Performance, Safety and Scale  @JetBrainsTV
Uploaded October 2026 | Updated November 2026, 3 hours ago
OpenAI builds some of the world's most advanced AI systems on infrastructure operating at extraordinary scale, where even small performance gains can have an enormous impact. Rust helps engineers keep those systems fast, reliable and safe.

Join Orhun Parmaksız and Predrag Gruevski, a software engineer working on Frontier Systems at OpenAI, for a practical look at Rust engineering inside OpenAI. They'll discuss how Rust helps solve demanding infrastructure challenges and how engineers balance speed, safety and reliability at scale. Predrag will also share how Codex is changing the development process by helping engineers explore ideas, test solutions and review their work more thoroughly.

Tune in to hear how Rust fits into OpenAI's work and what Predrag has learned from building systems at this scale.

00:00:00 Welcome and introductions
00:02:38 How did Predrag get into Rust?
00:04:36 The Rust a-ha moment
00:06:31 Trustfall and cargo-semver-checks
00:11:00 Bringing cargo-semver-checks into Rust's CI
00:17:05 Handling standard-library edge cases with Trustfall
00:20:18 How Predrag's use of AI has evolved
00:25:39 Using AI to harden GitHub Actions
00:27:16 When can you trust code you haven't fully read?
00:34:06 Do you still write code by hand?
00:36:24 Optimizing the Rust compiler with Codex
00:43:16 Is this vibe coding?
00:48:39 Tips writing reliable Rust with AI
00:54:51 What does a Frontier Systems engineer do at OpenAI?
00:58:44 Faster GPU training recovery with Rust
01:03:23 From prototype to carefully reviewed unsafe Rust
01:09:11 Building an AI skill for reviewing unsafe Rust
01:17:07 The cargo-miri vulnerability
01:21:47 AI security findings
01:25:31 AI-assisted Rust rewrites
01:28:25 Incremental Python-to-Rust migrations with PyO3
01:32:05 How OpenAI supports the Rust ecosystem
01:37:56 Favorite Codex features
01:40:09 Using Codex outside programming
01:42:37 What's next for OpenAI and Codex?
01:45:00 The future of Rust in an AI-powered world
01:47:36 Closing advice
01:50:21 Wrap-up
Rust at OpenAI: Performance, Safety and ScaleWhy Python is called Python?You don’t have to choose between Claude Code and RiderHow AI Is Changing the Way We Structure Engineering Work – and How It Shouldn’tAir inside Rider: a live demo from .NET DayAir Is Now in Every JetBrains IDE. Here’s Why.New NuGet Window in Rider🎬IntelliJ IDEA: The Documentary is now live on the CultRepo channel!The Art of Strategy | Meet JetBrains VP of Strategy Anastasia ZemskovaHow to Critically Evaluate AI-Written Code?File-based C# apps in RiderKotlin 2026: Layoffs, AI, Google — Is the Golden Age Over? Jake Wharton Explains
JetBrains |

Rust at OpenAI: Performance, Safety and Scale

SHARE TO X SHARE TO REDDIT SHARE TO FACEBOOK WALLPAPER