Flight GPU Engineer for Spacecraft HPC & AI

Doist; San Jose (CA); Hybrid; USD 156,000 - 186,000; Full time

Job summary

Muon Space is seeking a GPU Engineer to join the High Performance Compute team. You will design GPU-accelerated software for onboard orbiting satellites, enabling Earth-imaging sensor processing, RF analysis, and ML inference.

Hybrid work requires three days on-site in San Jose, CA. You will Port and optimize code for onboard GPUs, collaborate with FPGA, flight software, and system hardware teams, and ensure real-time performance within power constraints.

Qualifications

• 3+ years of experience developing GPU-accelerated software.
• Proficiency in C/C++ or Rust and Python.
• Experience shipping GPU-accelerated software using CUDA, OpenCL, HIP, or similar.
• Deep knowledge of GPU architecture and memory hierarchies.
• Experience with embedded Linux and cross-compilation, including Jetson/IGX-class platforms.

Responsibilities

• Design and implement GPU compute kernels for onboard image, RF, and ML workloads.
• Architect end-to-end GPU pipelines for live sensor data ingestion and processing.
• Port, adapt, and optimize code and models for on-orbit execution under real-time constraints.
• Profile and optimize GPU workloads to meet real-time deadlines under power budgets.
• Own verification, tests, and CI/CD for GPU software; ensure flight-readiness.

Skills

GPU programming

CUDA/OpenCL/HIP

C/C++

Python

Embedded Linux

Cross-compilation

CI/CD

Communication

Security clearance

Education

Bachelor’s or Master’s degree in CS/CE/EE

Tools

Docker

Cross-compilation toolchains

Linux/Jetson platforms

Job description

Muon Space is seeking a GPU Engineer to join the High Performance Compute team. You will design GPU-accelerated software for onboard orbiting satellites, enabling Earth-imaging sensor processing, RF analysis, and ML inference.

Hybrid work requires three days on-site in San Jose, CA. You will Port and optimize code for onboard GPUs, collaborate with FPGA, flight software, and system hardware teams, and ensure real-time performance within power constraints.