Boson is a Silicon Valley-based start-up specializing in farm automation. We tackle agricultural challenges through innovative electric vehicles and advanced computer vision technology. As we expand our operations and impact we are seeking a seasoned Robotics Engineer to lead the design , deployment and refinement of autonomous navigation systems for our Ackermann steered robotic vehicles. This is a production focussed role for a candidate with a track record of shipping and maintaining real robots in the field. You will lead efforts in mapping, localization, planner design, and perception using ROS2 / ROS1
Key Responsibilities
- ROS Navigation Stack : Design, implement, and maintain the ROS2 / ROS1 navigation stack (using Nav2 or equivalent), specifically tuned for Ackermann steering constraints (minimum turning radius)
- SLAM & Localization : Integrate SLAM / mapping and localization algorithms (e.g., SLAM toolbox, Cartographer, AMCL) to generate maps and provide robust state estimation
- Path Planning & Control : Develop global and local planners, cost-maps, and motion controllers (e.g., MPC, Pure Pursuit) that respect Ackermann kinematics and trajectory following
- Perception & Obstacle Avoidance : Fuse multisensor data (LIDAR, cameras, depth sensors, IMU, GPS / GNSS) for realtime detection, tracking, and prediction of static and dynamic obstacles, and building cost maps
- Computer Vision : Implement vision tasks such as object detection / recognition, semantic / instance segmentation, and depth estimation
- Robustness & Safety : Implement safety, fallback, and recovery behaviors for situations like being stuck, lost localization, or environmental anomalies
- Software Development : Write production-grade, real-time Modern C++ (C++17+) and / or Python code. Use product development practices including CI / CD and focus on performance optimization (latency, CPU / GPU)
- Team Leadership : Lead and mentor junior engineers and interns
Required Skills & Qualifications
The ideal candidate will have 7- 10 years of experience and deep expertise in the following :
Core Robotics & Software
ROS / Nav2 Expertise : Strong experience with ROS2 / ROS1 (Foxy, Humble, Jazzy, etc.), Nav2, and its design patterns (Lifecycle Nodes, Behavior Trees, Pluginlib, Composable Nodes)C++ / Python Proficiency : Expert in C++ and / or Python for real-time node development. Proficient in Modern C++ best practices (RAII, smart pointers, templates).Algorithmic Foundation : Thorough understanding of real-time programming, linear algebra, data structures, algorithms, kinematics, and 3D geometrySimulation & Testing : Experience with simulation (Gazebo, RViz, IsaacSim) and the ability to transfer solutions from sim to real hardware. Strong experience with ROS 2 unit, integration, and launch tests.Navigation & Perception Focus
Ackermann Kinematics : Good understanding of Ackermann steering constraints and how to adapt controllers / planners accordinglySLAM / Localization : Expertise in SLAM, localization, mapping, map representations, and transforms (tf2)Sensor Modalities : Proficiency with LIDAR, depth / RGB / Stereo cameras, IMUs, encoders, including calibration, synchronization, and noise filtering.Advanced Planning : Expertise in global / local planners, costmap generation, safety constraints, and designing hybrid planning frameworks.Perception : Strong skills in computer vision (object detection, segmentation) and the ability to integrate trained models (e.g., YOLO)Data handling skills
Machine Learning Frameworks : Proficiency in PyTorch, TensorFlow, and Keras for developing and deploying deep learning models in robotics applications.Data Handling : Experienced with Pandas for efficient data manipulation and preprocessing tasks.GPU Programming : Strong background in CUDA programming to leverage GPU acceleration for computationally intensive tasks.Nice-to-Have Skills
Experience with d ynamic obstacle prediction .Prior work with hybrid planners ( e.g., Hybrid A, SmacPlanner) that respect curvature constraintsExperience with hardware-in-the-loop testing and tuning PID / MPC controllersFamiliarity with robotics safety standardsReinforcement Learning for Deliberation in ROS 2Skills Required
lidar, gnss , Gps, Cuda, Tensorflow, Pandas, Pytorch, Keras, Python, Computer Vision