GPU Software Specialist, Onboard Compute

Muon Space•San Jose, CA

4d•$156,000 - $186,000•Hybrid

About The Position

Muon Space is seeking a GPU Engineer to join our High Performance Compute (HPC) team. You will design and develop GPU-accelerated software that runs onboard orbiting satellites, powering mission-critical workloads such as Earth-imaging sensor processing, RF signal analysis, and onboard AI/ML inference. On our satellites, GPUs will support mission-critical functions including Earth imaging sensor analysis, radio signal analysis, and high-performance onboard compute acceleration. You'll work across the full development lifecycle: feasibility, concept, architecture, design, implementation, verification, lab qualification, and deployment to flight, collaborating closely with FPGA engineers, flight software, payload and system hardware, and mission operations teams. This position is hybrid and requires working on-site in our San Jose, CA office three days per week.

Requirements

Bachelor's or Master’s degree in Computer Science, Computer Engineering, Electrical Engineering, Applied Math, or a related technical field, plus 3+ years of professional experience developing GPU-accelerated software.
Strong proficiency in C/C++ or Rust, and Python, with deep familiarity with memory management, concurrency, and performance-oriented programming in production systems.
Demonstrated production experience shipping GPU-accelerated software using one or more of CUDA, OpenCL, HIP, or similar.
Deep working knowledge of GPU architecture including SIMT execution, shared vs. global memory hierarchies, memory access patterns, occupancy, kernel launch overhead, and the tradeoffs between them when tuning real workloads.
Proven experience developing and debugging on embedded Linux (e.g., Ubuntu on Nvidia Jetson/IGX-class platforms), including cross-compilation, device tree basics, and userspace/kernel driver interaction.
Ability to write Linux userspace software integrating GPU compute with the rest of the system via shared memory or similar mechanisms.
Ability to work directly with internal and external customers, understanding their algorithms and models, and guiding them through the process of adapting those workloads to run efficiently on embedded GPU hardware.
Strong written and verbal communication skills, with the ability to produce clear design documents, interface descriptions, and test reports, and to lead technical discussions with cross-functional stakeholders.
Ability and willingness to obtain and maintain a U.S. security clearance. Active clearance is a plus.

Nice To Haves

Production experience with raw image or radio signal processing on GPUs.
Experience deploying ML inference on GPUs, including quantization and model optimization for edge deployment.
Hands-on experience with DSP/RF workloads on GPUs.
Hands-on experience with high-throughput data movement, multi-stream Ethernet-based payload streams and RDMA, including zero-copy buffer techniques.
Experience defining and operating CI/CD for embedded software and containerization, including hardware-in-the-loop test automation.
Direct exposure to space, aerospace, or other mission-critical software environments.
Prior experience in a customer-facing or applied-engineering role (e.g., solutions engineer, developer-relations engineer, or applied ML/DSP engineer), translating customer algorithms and models onto accelerator hardware.

Responsibilities

Design and implement GPU compute kernels (e.g. CUDA) for onboard image processing, radio signal processing, and ML inference workloads.
Architect end-to-end GPU pipelines that ingest live sensor data (optical and radio), process it on GPU, and hand results off to downlink or CPU-based decision-making subsystems.
Partner with internal and external customers to transform their algorithmic needs and inference models into efficient, flight-ready implementations on our GPU platform, including porting, adapting, and optimizing the code, models, and reference algorithms for on-orbit execution under real-time and power constraints.
Profile and optimize GPU workloads; tune occupancy, kernel launch configurations, and memory access patterns to meet real-time deadlines under strict power budgets.
Own verification and validation of GPU software, including unit tests, system-level tests, and hardware-in-the-loop tests for GPU software.
Define, evolve and own the GPU build and CI/CD environment, including cross-compilation toolchains for embedded targets and containerized builds.
Collaborate with flight software, FPGA, and payload and system hardware engineers on interfaces, data formats, and timing budgets between GPU compute blocks and the rest of the spacecraft.
Translate mission and payload requirements into robust, well-documented designs, trade studies, and interface documentation.