What You’ll Do
Own the end-to-end delivery of a GPU/SoM carrier board: requirements, architecture, schematics, layout guidance, prototypes, bring-up, validation, and production support
Design and integrate high-speed interfaces: MIPI CSI-2, PCIe, USB 3.x, Ethernet, DDR (SoM-side coordination), I2C/SPI/UART, CAN
Interface with flight controller systems (e.g., Pixhawk-class) over CAN/UART/I2C and ensure electrical robustness in real-world environments
Integrate wireless modules (Wi-Fi/LTE/5G/private radio): power, PCIe/USB/UART connectivity, antenna/RF coordination, EMI/EMC risk control
Build power architecture: rail budgeting, sequencing, protection (OVP/OCP/ESD), transient/thermal considerations, and margin analysis
Create validation plans and verify features and interfaces across corners (voltage, temperature, load, EMI)
Drive DFM/DFT/EMC readiness: test points, manufacturing constraints, rework strategy, and compliance preparation
Produce clear documentation: interface specs, power budgets, design rationale, test plans, and issue tracking/closure
Job Requirements
B.S. in Electrical Engineering or equivalent practical experience
2+ years of hands-on PCB/PCBA design with full bring-up and debug ownership
Strong high-speed digital design experience (>5 Gbps) and interface know-how: PCIe, MIPI, USB3, Ethernet, DDR
Solid understanding of power and analog basics: switching regulators, LDOs, MOSFETs, op-amps, protection circuits
Proficient with lab tools: oscilloscope, logic analyzer, protocol analyzers, signal generators (BERT/RF tools a plus)
Experience supporting EVT/DVT/PVT and working with manufacturers on yield and reliability
Strong cross-functional collaboration with layout, firmware, mechanical, EMC, and program teams
Preferred
Jetson/Xavier/Orin/AGX (or similar GPU SoM) carrier board experience
Multi-camera synchronization / hardware timestamping / PTP experience
UAV/robotics system integration experience (flight controller, CAN robustness, field reliability)
AI/Transformer knowledge is not required, but understanding compute and sensor bandwidth trade-offs is a plus
