Really interesting perspective on how Rust, AI, and engineering leadership intersect. How do you think Rust will shape the future of AI engineering?
Mike Dabydeen on Rust, AI & Engineering Leadership
16 Comments
@[Mehadi Hasan] Mostly in the layer around the model, not the model itself. Python will stay where training and research happen. Rust is showing up where the work runs over and over: Hugging Face's tokenizers library has a Rust core, and OpenAI's Codex CLI ships a Rust implementation, codex-rs. That's the same place I care about, agent harnesses and tool servers that run on small local hardware, where memory safety and low overhead decide whether it's usable. I expect more of that layer to move to Rust as people run agents locally and start paying attention to cost per run.
Please log in to add a comment.
@[Faisal Ashraf] Thank you, Faisal. That means a lot. Helping developers learn and grow is the part of this work I enjoy most, whether it's in a classroom at Sheridan or Conestoga or in threads like this one. If anything from the interview connects to something you're working on, I'm happy to help, so feel free to ask here anytime.
Please log in to add a comment.
it was a really eye opening interview, I appreciate the part where you encouraged system thinking because as software scales higher it becomes difficult to reap benefit if llm are given full automatic access since they can't handle scaling without expert supervision.
it's good to hear about rust ability to handle memory issues using the borrow checker, especially in this days where manual memory management can be difficult and buggy.
Thank you for sharing those amazing insights on current issues baffling new developers and existing ones.
@[peter isaiah] Thanks, Peter. The borrow checker is a good model for the supervision part too. It doesn't depend on a senior engineer spotting a use-after-move in review; rustc rejects it with E0382 on every build, however confident the author was. With agents I'd aim for the same thing: put the limits that must hold in checks that run every time, and save the expert's attention for what a check can't decide, like whether the design is right in the first place.
Please log in to add a comment.
@[James Dayal] Thank you, James, and a big thank you to Tom Smith for hosting such a thoughtful conversation. I really appreciate you and the CoderLegion team giving developers a place to share their stories and learn from each other. It was a pleasure to be part of the series, and I'm grateful for the warm welcome everyone here has given it.
Please log in to add a comment.
@[Sascha] Thank you, Sascha, I really appreciate that. You put it well. Technical skill gets you part of the way, but the judgment about what to build, and when to slow down and check, matters just as much, and it keeps shifting as the tools change. I'll keep sharing what I learn, and if you're working through any of this yourself, I'm always happy to compare notes.
Please log in to add a comment.
Please log in to comment on this post.
More Posts
- © 2026 Coder Legion
- Feedback / Bug
- Privacy
- About Us
- Contacts
- You Tube
- Tiktok
- Premium Subscription
- Terms of Service
- Early Builders
I donβt just write codeβI break down complex ideas, evaluate tools, and contribute perspectives that help developers make better technical decisions. Show less
More From James Dayalverified
Related Jobs
- Senior DevOps Engineer - Cloud Platform EngineeringUST Global Β· Full time Β· Norfolk, VA
- Technical Program Manager (Engineering Project Lead)Seek Thermal Β· Full time Β· Puerto Rico
- Software Engineering Co-Op (Summer/Fall 2027) (Hybrid)RTX Β· Internship Β· Puerto Rico
Commenters (This Week)
Contribute meaningful comments to climb the leaderboard and earn badges!