Mastering Rust on Edge and Cloud — AI/IoT Case Studies
summary
This talk describes how AI Drive Chart rewrote a Python R&D pipeline into Rust for both edge devices and cloud, delivering a working prototype in two weeks and an incremental migration to production that has shown no crash-causing memory bugs. The team used Cargo, GitHub Actions, dependabot, snapshot tests (Intsa), bindgen/CPP for C/C++ interop, and exposed inference via gRPC/Tonic or a local C API while logging with s-log JSON traces; a four‑person team completed refactoring into micro‑crates over about six months.
AI-written summary · about
chapters (13)
- 0:00Start
- 2:03Rust in DRIVE CHART
- 2:37Why Rust?
- 5:12Development Flow of Edge AI Feature
- 6:49Rust in Drive Recorder
- 8:15How To Avoid Degradation
- 9:36Rust In Cloud
- 10:59Internal Structure of Edge AI Crate
- 12:22Implementation of Hexagonal Architecture in Rust (1/2)
- 14:12Debugging, Tracing, and Profiling
- 15:22How the System Has Changed over the Past Year
- 16:54How the Team Has Changed over the Past Year
- 18:36Conclusion
related talks
19mEnhancing Gen-AI with Rust: Scalable AI Systems
01h 00mFrom Tiny Chips to Server Racks: Deploying AI Everywhere with Rust
23mRust in Edge Computing: Enhance Performance & Security
27mHow to (not) introduce Rust at your workplace - a tale
20m"Rust in Motion: Building Reliable and Performant Robotics Systems"
04mOxidize Lightning Talks – Friedrich Rockenbauer
21mRust - Cargo Cult?
01h 27mRust in Academia and Introduction to Iron