Research Engineer

At XDOF, we’re at an inflection point. Frontier labs are racing to build general-purpose robots, and high-quality training data is the bottleneck. We’re building the foundation behind the foundation models – the data collection systems, operational capability, exabyte-scale data warehouse, and software toolchain – to help our partners drive the field forward.

Our research teams move fast and produce breakthrough work, but research code and production code are different things. We’re looking for a Research Engineer to bridge that gap: someone who can read a research prototype, understand it deeply, and turn it into something that runs reliably at scale on real hardware. You can expect to float across teams to wherever the highest-priority needs are, across perception, ML, and data infrastructure.

What You’ll Do

Research engineers take prototype code and make it production-grade. Sample projects include:

taking a research perception pipeline (pose estimation, SLAM, calibration) and hardening it for reliable, real-time execution on embedded platforms

profiling and optimizing performance-critical code at the CPU, memory, and GPU level using tools like perf, NSight, and custom microbenchmarks

writing and debugging CUDA kernels for low-level acceleration of compute-heavy workloads

integrating research outputs into the production codebase with proper testing, error handling, and observability

containerizing and packaging workloads (Docker) so they can be scaled and deployed by the infrastructure team

understanding and leveraging the infrastructure team’s orchestration and compute systems to hand off production-ready workloads cleanly

working with researchers to understand algorithmic intent and make informed tradeoffs between accuracy, latency, and resource usage

About You

Baseline skills

3+ years of industry experience in software engineering with a focus on systems, performance, or production ML

strong C++ proficiency, including modern C++ (C++17/20), memory management, and performance-conscious coding patterns

CUDA programming experience: ability to write, profile, and debug GPU kernels

experience with CPU performance optimization: profiling, cache behavior, SIMD, latency reduction

proficiency with Python and familiarity with ML frameworks (PyTorch, TensorFlow) at the level needed to read and modify research code

comfort with Linux systems, including build systems, debugging tools, and containerization

You might be a good fit if you

have taken research or prototype code and shipped it in a production system

have worked on real-time or embedded systems where latency and resource constraints matter

have experience with perception, computer vision, or robotics systems

have optimized model inference for deployment (TensorRT, ONNX Runtime, or similar)

understand the full lifecycle from research notebook to containerized, monitored production service

are very comfortable working in 0→1 environments

are mission-driven and passionate about robotics: work at XDOF is fast-paced and constant. We hope you love what you’re going to be doing, because you’ll be doing a lot of it!