You will own the middleware integration layer for the 4NE-1 Humanoid Software Platform, ensuring reliable communication between perception, planning, locomotion, whole-body control, manipulation, state estimation, hardware interfaces, and fleet-level services. The humanoid stack spans multiple middleware paradigms, ranging from hard real-time control loops to distributed cloud-connected robotics systems, all of which must operate seamlessly on a production humanoid robot.
Your Mission & Challenges
- ROS 2 Platform Integration: own the ROS 2 infrastructure used across the 4NE-1 software stack, including sensor and actuator interfaces, state estimation pipelines, perception systems, locomotion, manipulation, diagnostics, and fleet connectivity. Responsibilities include ROS 2 architecture, driver integration,
ros2_control configuration, DDS tuning, launch system design, diagnostics, and migration across ROS 2 releases such as Humble and Jazzy. - OROCOS RTT Real-Time Integration: own the hard real-time middleware layer that powers humanoid locomotion, whole-body control, inverse dynamics, and low-level control pipelines. This includes OROCOS RTT component architecture, deterministic execution, inter-process communication, and service abstractions used to bridge high-level software with safety-critical hardware interfaces.
- Apex.OS & Functional Safety Deployment: develop the path from research-oriented ROS 2 software to safety-ready production deployments. This includes integration of Apex.OS and Apex.Middleware where required, supporting deterministic execution and future safety-certifiable humanoid applications that must meet functional safety requirements.
- DDS Communication Architecture: design and maintain the DDS communication infrastructure used throughout the humanoid platform. Integrate, configure, and optimize FastDDS, Cyclone DDS, and RTI Connext deployments, ensuring reliable communication between distributed software components operating under strict timing constraints. Understand and troubleshoot reliability, discovery, latency, fragmentation, and QoS behavior at the RTPS protocol level.
- Zenoh & Distributed Robot Connectivity: drive the integration of Eclipse Zenoh for distributed robotics use cases, including remote operation, fleet supervision, cloud connectivity, teleoperation, data collection, and multi-robot communication. Ensure reliable interoperability with existing ROS 2 and DDS-based systems.
- Zero-Copy & High-Bandwidth Data Transport: develop low-latency data transport mechanisms for bandwidth-intensive humanoid workloads such as perception pipelines, whole-body state streaming, motion planning, simulation interfaces, and real-time control. Leverage technologies such as Iceoryx2 and shared-memory DDS transports to eliminate unnecessary copies and minimize communication latency.
- Cross-Stack System Integration: ensure consistent communication contracts across the complete humanoid software stack. State estimation, perception, motion planning, manipulation, locomotion, whole-body control, simulation, and hardware interfaces must share coherent message definitions, time synchronization strategies, coordinate frame conventions, and interface standards as the robot evolves.
- Real-Time Performance & Reliability: work closely with locomotion, controls, perception, and systems engineers to identify communication bottlenecks, deterministic execution issues, synchronization problems, and latency-related failures. Enable reliable operation of control pipelines running from high-frequency actuator loops up to fleet-level software services.
What we can look forward to
- Strong C++ expertise with real-time-safe software development practices, including lock-free programming, memory management, zero-copy communication, and deterministic execution.
- Deep understanding of Linux-based robotic systems and real-time middleware architectures.
- Practical experience integrating middleware into real-time robotics systems operating under strict latency and determinism requirements.
- Strong understanding of DDS concepts, including QoS policies, discovery mechanisms, reliability models, durability, history management, fragmentation, and network-level performance tuning.
- Production experience in at least one of the following domains:
ROS 2 architecture, ros2_control, lifecycle management, hardware interfaces, and system integration.
- OROCOS RTT and hard real-time robotics software.
- Apex.OS / Apex.Middleware deployments in industrial or safety-critical environments.
- DDS infrastructure development using FastDDS, Cyclone DDS, or RTI Connext.
- Experience collaborating with controls, locomotion, manipulation, perception, and embedded software teams in complex robotic systems.
- Solid understanding of networking, Linux performance tuning, real-time kernels, profiling, debugging, and distributed systems.
Nice to Have
- Experience developing software for humanoid robots, legged robots, or other complex floating-base robotic systems.
- Hands-on experience supporting locomotion, whole-body control, manipulation, perception, or state estimation pipelines.
- Experience integrating communication frameworks into high-frequency control loops (500 Hz–1 kHz).
- Eclipse Zenoh for distributed robotics, teleoperation, or fleet-management applications.
- Iceoryx or other zero-copy IPC technologies.
- Experience bridging multiple middleware ecosystems, including ROS 2, OROCOS RTT, DDS, and custom robotics frameworks.
- DDS Security, authentication, encryption, and access-control mechanisms.
- Contributions to ROS 2, OROCOS, Zenoh, DDS implementations, or other robotics infrastructure projects.
- Experience working on large-scale robotics software platforms with dozens of interacting subsystems and multidisciplinary teams.
