Bernardo Costa is a systems-focused software engineer with 11 years of experience building and maintaining Rust and Nix-based build infrastructure at scale, currently driving Rust/Nix build tooling at Anthropic. He led the Rust build system team at AWS, shipping improvements that serve the entire company, and has applied that expertise to compiler and performance tooling contributions in the rust-lang/rust repo. A longtime NixOS maintainer and contributor to projects like home-manager and cachix, he blends devops craftsmanship with language- and build-system engineering. Earlier work spans quantum control and compiler tooling at Google Quantum AI and productionizing high-throughput camera and data pipelines in Rust at Standard AI. He’s comfortable moving between low-level firmware and large distributed build systems, and has a track record of replacing fragile Python pipelines with robust Rust services. Based in New York and trained in computer engineering at Instituto Superior Técnico, he pairs deep technical ownership with open-source stewardship.
11 years of coding experience
10 years of employment as a software developer
Computer Engineering, Computer Engineering at Instituto Superior Técnico
Manage a user environment using Nix [maintainer=@rycee]
Role in this project:
DevOps Engineer
Contributions:2 reviews, 4 commits, 19 PRs in 2 years 8 months
Contributions summary:Bernardo's contributions primarily focused on improving the build and deployment process and enabling home-manager functionalities. They implemented changes to integrate environment variables into systemd services within the Sway window manager. The user also added a flag to the home-manager tool and fixed issues related to caching of configuration and dependencies. Furthermore, the user removed deprecated options and made changes to allow multiple keybindings for history substring search in ZSH.
Contributions:12 commits, 3 PRs, 33 comments in 6 months
Contributions summary:Bernardo primarily focused on parallelizing and optimizing the `absubmit` plugin within the beets music library manager. They initially implemented parallel processing using `multiprocessing.Pool` and subsequently transitioned to `multiprocessing.ThreadPool` and `concurrent.futures.ThreadPoolExecutor` for improved performance and Python version compatibility. They also refactored the code to use a utility function `par_map` for parallel execution. Additionally, the user updated the changelog to reflect these performance improvements.
clipythonmusicbrainztaggermusic
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.
Request Free Trial
Bernardo Costa - Member Of Technical Staff at Anthropic