Member of Technical Staff – Senior Engineer, VR Teleoperation (Headset and Graphics)

Walden Robotics•Cambridge, MA

33d

About The Position

You will build and own major pieces of the headset side of Walden's teleoperation system—the immersive OpenXR application through which remote operators see, hear, feel and control our humanoid robots in real deployments. This is a generalist VR engineering role with a focus on graphics: you'll move comfortably across the OpenXR stack, head and hand tracking, input and frame timing, and bring the graphics depth to be the go-to engineer for how pixels get to the display. The robot's cameras are never where the operator's eyes are, and frames always arrive late; you'll hide that with prediction, reprojection and compositing that work with each headset's runtime. The work is driven by measurement: you'll shave latency, cut power consumption and raise framerate and resolution, with metrics to show each gain. You'll join a small, hands-on teleoperation team working to a shared architecture and latency budget, and thrive on quick iteration from idea to on-device prototype.

Requirements

Production Shipping Experience: Shipped real-time graphics or XR software in production, with working knowledge of compositors, timewarp and frame pacing (orthe ability to pick them up quickly).
3D Math and Tracking: Deep, practical command of projective geometry, homographies, camera models, coordinate frames, SE(3) transforms, quaternions and pose tracking. You can derive a warp, not only call one.
Real-Time and Mobile GPU Graphics: Strong Vulkan or OpenGL ES and shader experience, including profiling and optimizing rendering for mobile GPUs across bandwidth, overdraw and power.
OpenXR and VR Fluency: Comfortable working across the OpenXR stack and VR application code: sessions, spaces, tracking, swapchains and composition layers.
Fast, Solid C++: Strong modern C++ on the real-time path and Python for tooling, with a track record of turning fast prototypes into reliable software.
AI-Assisted Development: Demonstrated use of AI coding tools (for example Claude Code, Cursor or Copilot) to substantially increase your productivity, with the judgment to review, test and own that code on a real-time path.
Cross-Disciplinary Collaboration: Comfort landing features with robotics, controls and networking engineers that span the headset and the robot.
End-to-End Ownership: Demonstrated ability to take features from design through deployment and field validation with minimal supervision.

Nice To Haves

Native OpenXR development in C++ without a game engine, including direct Vulkan swapchain and composition-layer management; Rust is a plus.
Optimization for Qualcomm Adreno tile-based rendering (GMEM usage, binning, fragment density maps).
Shipping across multiple OpenXR runtimes (for example Meta, Android XR, Pico and SteamVR) and handling their differences in extensions and compositor behavior.
Asynchronous or application spacewarp-style reprojection, or deep knowledge of how headset runtimes and compositors reproject.
Image-based or remote rendering, including view synthesis from wide-field-of-view cameras.
Hardware video decode paths (MediaCodec, AHardwareBuffer) and zero-copy decode-to-texture.
Hands-on profiling with Perfetto, RenderDoc and simpleperf (Snapdragon Profiler or OVR Metrics Tool a plus).
Networking for real-time streaming (WebRTC, QUIC or custom transport) and familiarity with WiFi streaming challenges such as jitter, loss, bandwidth variability and roaming.
Familiarity with Unity/C# on standalone headsets.
Track record of leading the technical design of major XR or graphics subsystems at scale, mentoring engineers and driving technical consensus across teams.

Responsibilities

Build, maintain and improve the core of the operator application beneath the UI—frame loop, streaming client, rendering and platform layer—and evolve its architecture as the product matures, debugging issues in real deployments.
Build on each headset runtime's head, hand and controller tracking through OpenXR, adapting it for the full loop from the operator's motion to the robot and back to the display, and return haptic feedback to the operator.
Implement pose prediction, frame timing, reprojection and compositing that work with each headset runtime to keep the operator's view stable and responsive despite network and rendering delays.
Profile and optimize rendering on Snapdragon-based XR headsets, including tile-based rendering and foveation, to hold framerate at minimal power.
Instrument latency, power, framerate and GPU cost, and use the data to cut latency and power while raising framerate and resolution.
Keep the application working across headsets (for example Meta, Android XR, Pico and Steam), and bring it up on new ones as they are evaluated.
Partner with robot-side teleoperation, networking and UX engineers to define interfaces and land features end-to-end.