# Ryan Butler > Software Engineer at Worldcoin Location: Brooklyn, New York, United States Profile: https://flows.cv/ryanbutler Experienced Rust developer passionate about Virtual Reality, realtime systems, and open source. Let's build something awesome, together :) ## Work Experience ### Software Engineer @ World Jan 2023 – Present | New York, NY • Implemented binary diffing in our over-the-air update mechanism, reducing OTA sizes by 150x (6GB -> 40MB). • Wrote Rust OP-TEE/TrustZone code to handle secure encrypted storage, with novel access controls and user isolation. • Rapid prototypes: peer to peer blob storage via iroh/QUIC, gstreamer gpu codecs with webtransport video streaming, etc. • Singularly responsible for all open sourcing in the company, including directly coordinating with internal and external legal counsel, security teams, engineers, and the technical challenges of balancing transparency with security. • Sole maintainer of a NixOS hardware-in-loop tester, with custom Rust tooling for booting Nvidia Jetsons via RCM + NFS boots, scripting FTDI gpio to automate physical buttons, serial tty processing, and stm32 readout protection in probe-rs. • Maintained most Rust systemd services and all of my team’s Rust tooling, CI, build infrastructure, docs, and DevX. • Low level userspace linux - DBus (via zbus crate), systemd, dm_verity, dm_crypt, deb packaging, NixOS, etc. • Implemented support for the orb’s thermal camera. Wrote safe FFI Rust <-> C bindings, with async tokio interop. ### Open Source Maintainer (Paid) @ SlimeVR Full-Body Tracking Jan 2022 – Jan 2023 • Built bare-metal and async Rust firmware for full body tracking. Improved power efficiency by 5x and latency by 3x over Arduino. Designed a general framework for sensor fusion, networking, and IMUs, for different combinations of hardware. • Shipped software to hundreds of users, allowing them to view their pose in 3D, overlaid on top of other VR games. • Primary author of SolarXR, a realtime messaging protocol using Flatbuffers and Websockets. Enabled a transition to a polyglot codebase and eliminated thousands of lines of code as the team adopted Tauri, React, and TypeScript. • Created a small scale business selling custom wearable circuit boards I designed in KiCad for full-body tracking. ### Lead Software Engineer (Platform) @ Teleportal Jan 2021 – Jan 2023 | New York, New York, United States • Lead designer and author of an open source, realtime data model for Virtual and Augmented Reality written in Rust. • Wrote Rust macros to allow user defined types and dynamic types to be stored in the data model. • Created a novel solution to allow Rust to be called safely from C#, and deployed Rust code inside of a Unity app. • Implemented full-body motion capture on standalone Quest using IMUs, Websockets, Rust, and Unity, on the Quest pro. • Built a job runner backend in Rust using the Axum web framework, HTTP, Websockets, Tokio, Tracing, and Docker. ### Software Engineer @ Samsung Artificial Intelligence Center Jan 2019 – Jan 2021 | New York, United States • Built a high performance data pipeline and event based exponential filter in Rust, improving performance by ~750x. • Implemented a high performance messaging framework using Flatbuffers and ZeroMQ in C++ and Rust. Added support for many different hardware devices and sensors. Integrated the framework with ROS2, a popular robotics toolkit. • Implemented a low latency, zero-allocation event decoder in Rust for a neuromorphic event based camera (DVS). • Open sourced work improving Bazel’s support for creating Python bindings for Rust and C++ (Pybind11 & PyO3). • Singularly responsible for teamwide transition to the Bazel build system, and improving Gitlab CI infrastructure. • Taught several workshops on type systems, Rust, and the tradeoffs in programming language design and performance. ### Software Engineering Intern, Think Tank Team @ Samsung Research America Jan 2019 – Jan 2019 | Mountain View, California • Worked in Samsung's special projects group in machine learning and software engineering. • Built a neural text to speech engine in Python using PyTorch, TensorFlow, WaveNet, Tacotron 2, and WaveRNN. • Implemented multi-GPU and device-agnostic training of Tacotron and WaveRNN in PyTorch. • Designed data scraping methods using Python, Pandas, and Google Cloud APIs. ### Deep Learning Team Cofounder, Executive Board @ Cornell Data Science Jan 2017 – Jan 2018 | Cornell University, Ithaca, NY • Cofounded a team of ~20, designed their training program, and directed their research. • Built a deep Q-learning network to predict power grid distribution. Built and optimized a custom C++11 engine to simulate population growth and power distribution. Interfaced between the C++ environment and Python reinforcement learning agent. • Built a Reinforcement Learning library for making state of the art, highly parallel algorithms like A2C easy to use. Designed a vectorized API to run OpenAI Gym environments in parallel. • Lead instructor and creator of “Deep Learning in TensorFlow” lecture series. Taught over 200 graduate and undergraduate students the theory and implementation of fundamental neural networks in TensorFlow • Maintained ~$8000 worth of Linux servers and built a ~$4000 GPU server for training deep learning models. ### Undergraduate Researcher, Skynet Autonomous Car Group @ Autonomous Systems Lab Jan 2016 – Jan 2018 | Cornell University • Used novel neural attention mechanisms for point sets to create a neural network architecture to classify LIDAR data, with performance equivalent to state of the art methods. • Worked on autonomous vehicle perception. Created a convolutional encoder-decoder architecture for neural networks to semantically segment LIDAR data using TensorFlow. • Increased execution speeds by 25x by developing optimized C++ code to cluster LIDAR data. • Built a TensorFlow library to generalize common neural networks to an arbitrarily high number of spatial dimensions. • Received a research grant from the Semiconductor Research Corporation with Support from Intel Foundation. ### Software Engineering Intern, Systems & Architecture for Machine Learning @ IBM Jan 2017 – Jan 2017 | Watson Research Center, Yorktown Heights, NY • Invented a method to improve training speeds by 2x for any RNN in TensorFlow. • Researched methods to improve recurrent neural network architectures, specifically the Differentiable Neural Computer and the Neural Turing Machine. • Developed and tested multiple implementations of the backpropagation through time (BPTT) algorithm in TensorFlow on LSTMs and optimized their performance on GPUs and CPUs. ### Congressional Intern @ US House of Representatives, TX-27th District Jan 2016 – Jan 2016 | Corpus Christi, Texas • Assisted constituents by providing a professional and courteous point of contact. • Streamlined office productivity by implementing changes to the way constituents’ documents were digitally stored. ## Education ### Bachelor of Science - BS in Computer Science Cornell Duffield College of Engineering Jan 2015 – Jan 2019 ## Contact & Social - LinkedIn: https://linkedin.com/in/ryan-butler --- Source: https://flows.cv/ryanbutler JSON Resume: https://flows.cv/ryanbutler/resume.json Last updated: 2026-03-23