Rust engineering
Given the choice, I write Rust. It’s fast, it’s safe, and it does more with less hardware.
Rust gives you the performance of C or C++ without whole classes of memory and concurrency bugs. Programs start quickly, use little memory, and usually need smaller, cheaper servers, which also means less energy.
Where it pays off
- Performance-critical components: the hot path of a system written in C#, Go, TypeScript or Python, rewritten in Rust and called directly from the existing code.
- Services and APIs that need to be fast and predictable under load.
- Command-line tools and single-binary applications that are easy to distribute and run anywhere.
- Reliability: software that has to keep running, where a crash or a data race is expensive.
Experience
- Microsoft: in a high-security government delivery team, built production services in C# with performance-critical components in Rust and Go. I also built and benchmarked equivalent implementations in Python, Go, Rust and C#, called from C# through FFI, to compare their performance and trade-offs.
- An AI app-building platform: designed and built a local-first migration agent as a Rust workspace of around 18 crates, with a native desktop app built with Tauri, macOS screen capture and Keychain integration, local OCR, a CLI and a local MCP server. Its security guarantees are enforced by automated tests, not just stated as policy.
- Rilbo: our own project tracker is a single Rust executable, with the web UI, API and MCP server built in and the data in SQLite.
When Rust isn’t the answer
If your team works in TypeScript or C# and the bottleneck isn’t the language, a rewrite won’t help. I’ll say so. I also work in Go, TypeScript and C#/.NET, and I’ll recommend whichever your team can maintain.