| |
Rust Meetup - Bevy Rendering & Build Times
|
| With Chris Dryden (Dev Productivity, Amazon), Charles Giguere (Foresight Spatial Labs), Lawrence Harvey (Rust). |
| Materialize, 436 Lafayette, 6th Fl |
|
Feb 03 (Tue) , 2026 @ 06:30 PM
| |
FREE |
|
|
|
|
|
|
|
|
| |
| DETAILS |
|
Join us on Tuesday, February 3rd at Materialize. Doors open at 6:30 p.m. to give attendees plenty of time to grab pizza & socialize, & the talk begin at 7:15 p.m. Following the success of the UnConf we've absorbed all of your feedback & have two awesome speakers!
Charles Giguere is a Bevy SME at Foresight Spatial Labs. Modularity in Practice: How Bevy Powers Foresight Spatial Labs' renderer
Foresight Spatial Labs builds realtime spatial visualization tools for large, spatial datasets. We chose Bevy for its modern ECS architecture & highly modular renderer, which lets us iterate quickly, share improvements upstream, & maintain productiongrade customizations without forking the entire engine. This talk shares concrete engineering experiences: upstreaming orderindependent transparency (OIT) & picking improvements after starting with private implementations; developing GPUaccelerated picking that is planned for upstream contribution; & selectively replacing Bevy subsystems while keeping the rest of the engine intact. This project was actually upstreamed into Bevy itself.
Attendees will see how Bevy's modularity reduced timetoprototype, improved code quality through community review, & enabled a high-quality production renderer that meets domainspecific constraints.
Key takeaways
* A concrete pattern for using Bevy modularity: prototype locally, upstream what generalizes, & maintain small, welldocumented custom extensions.
* Case studies: OIT & CPU/GPU picking - from private feature to community contribution.
* How selective replacement (custom transforms, culling, render passes) interoperates with Bevy's renderer & ECS.
* How contributing back benefits both product robustness & the wider opensource ecosystem.
Chris Dryden works on Developer Productivity at Amazon.
How Amazon was able to reduce the average build time for all developers by 40% with the help of Rust.
Are you interested in leveraging Rust to improve your existing services & build processes? This presentation is about the journey that Amazon took to rebuild large parts of Amazon's build system in Rust & the lessons it learned along the way. Amazon was able to reduce the average build times for all developers by 40% in 2025. The presentation will go in depth into the tools used & what are the exact steps that you can take to identify what are the best candidates for rewriting & how to safely migrate critical systems relied on by a multi-trillion dollar company. This presentation will go in depth on how to leverage profiling tools such as Samply, & strategies on how to incrementally move components to Rust with monitoring & fallback mechanisms to make the migration process seamless. The presentation will also go through the lessons learned about what were the benefits & drawbacks of the large scale migrations & how you can set yourself up for success for your next migration at your own company.
This talk will:
Go through the background of the build system used by Amazon, & some background about how Rust is being adopted across the company in critical areas
Go through how we set up a telemetry & monitoring system before starting to make sure we know what we're doing doesn't impact existing workflows & measure how successful we can be
How to setup samply with your existing code & services to get a sense of where the best areas are to tackle & do migrations first
Leveraging AI to help with Fuzzing & setting up integration tests to make sure when doing the rewrite you don't have any regressiong
Lessons learned about where the best ROI comes from when doing these migrations & heuristics on where to focus your effort for the best return on time
Lawrence Harvey is Rust NYC's official recruitment partner, with Ross providing support as a co-organizer & financial support.
The space is generously sponsored by our partner Materialize.
|
|
|
|
|
|
|
|
