Senior Edge Architect

• 8+ years of experience in embedded or edge systems for robotics, autonomous vehicles, or industrial machines.
• Hands-on experience with NVIDIA Jetson or IGX platforms; experience with the Thor generation is preferred.
• Experience with camera interfaces (GMSL, MIPI CSI), automotive Ethernet, CAN, and storage throughput.
• Experience developing and customizing the Linux BSP, kernel, and NVIDIA drivers (JetPack, L4T) in C++.
• Practical experience with CUDA, TensorRT, DeepStream, and Isaac ROS.
• Experience building production ROS 2 nodes, tuning DDS, using rosbag2 or MCAP logging, and maintaining performance under high sensor load.
• Proven experience integrating and synchronizing cameras, LiDAR, radar, and GNSS/IMU (PTP/PPS).

Nice to have:

• Experience with off-road, construction, mining, or agricultural machines.
• Sensor calibration expertise, including online monitoring of calibration health.
• Experience with edge data reduction, compression and encryption, and device management, OTA, and remote diagnostics across hundreds to thousands of units.
• Safety-related embedded development under ISO 26262 or ISO 13849.

Edge Hardware Architecture:

• Define the next-generation kit around NVIDIA Thor-class modules, including Jetson AGX Thor and IGX Thor, covering sensor I/O, camera and network interfaces, storage, power, and enclosure for dusty, high-vibration conditions and terabyte-scale data per machine.

Low-Level Software Stack:

• Own the NVIDIA Linux BSP, kernel and drivers, device tree, boot, and over-the-air updates.

Middleware & Common Driver Layer:

• Design middleware on ROS 2 Jazzy and Isaac ROS, lead the migration to ROS 2, and build a common driver and service layer to add or remove cameras, LiDARs, radars, GNSS, and IMUs without redesign.

Time Sync, Logging & Onboard Inference:

• Set up precise time synchronization across sensors (GNSS/UTC, PPS, PTP), high-throughput logging that never drops data, and onboard GPU inference for detection, segmentation, and fusion with event-triggered capture within compute and thermal limits.

Fleet Operations at Scale:

• Build remote enable/disable, health and diagnostics reporting, and secure, encrypted data handling as the data-collection fleet scales to 500+ machines, making kit health visible remotely without engineers going on site.

Architecture Decisions & Roadmap:

• Decide with the robotics team whether data logging and autonomy compute share hardware.
• Co-author the 2027 program charter and roadmap with the client, turning architecture into working software quickly alongside AI coding agents.

• Availability for pre-approved trips to client sites.