Robotics Software Systems Engineer

Aurelius Systems•San Francisco, CA

49d•Onsite

About The Position

Aurelius Systems is a VC-backed defense tech startup building autonomous, edge-deployed robotics systems using directed energy for counter-UAS. We build laser weapons to shoot down drones. We're a small team of ~10 engineers, former US military operators, and subject matter experts scaling America's directed energy dominance. The first cost effective, reliable and robust laser weapon system. Our namesake isn't an accident. Marcus Aurelius wrote about doing the work in front of you, every day, without excuses. Henry Ford didn't wait for permission to reinvent manufacturing. That's how we operate; small team, unreasonable output, no hiding behind the unachievable. In addition to our San Francisco lab/office, we opened our Detroit manufacturing hub and we field test weekly on our own 400-acre private range. If you are the kind of engineer who'd rather see your work on a range than in a lab, read on. You'll be the engineer who ensures every subsystem is working together as efficiently as possible. Our system is a stack of complex subsystems: sensing, computer vision, ML inference, controls, power, mechanical actuation. It is paramount that processing time and inter-process latency is optimized to an absolute minimum in order to defeat our small, fast, evasive UAS targets. The critical gap we're hiring for is real-time systems performance at the hardware boundary. You need to understand how software execution translates into physical system behavior; how latency accumulates across CPU, GPU, memory, and I/O; how bandwidth limits affect sensor pipelines; and what happens when your model runs at 120 FPS in isolation but drops frames once integrated with the full system. We need an engineer who thinks in terms of microseconds, memory bandwidth, cache behavior, and determinism. At ~10 engineers you won't be a cog in a software org. You'll own systems that go to the range. Tests either work or they don't, and it’s up to you to iterate and improve.

Requirements

3–6+ years in robotics software engineering with real hardware systems
Expert-level C++ (modern C++17/20) systems programming
Deep understanding of GPU memory models (global, shared, unified memory)
ARM + Linux systems development (cross-compiling, profiling, kernel-level awareness)
Performance optimization across CPU/GPU boundaries
Shared memory and lock-free architecture design
Real-time or near real-time pipeline optimization
High-throughput peripheral data ingestion (USB, PCIe, Ethernet)
Multithreaded systems and concurrency optimization

Nice To Haves

Jetson platform experience
DMA and zero-copy pipeline design
CUDA kernel optimization for throughput and latency
Video pipeline experience (OpenCV, GStreamer, Vimba/Pylon)
CoaXPress, USB3 Vision, or high-speed camera systems
Linux kernel or driver-level experience

Responsibilities

Own the latency budget of the full platform, from sensing through actuation
Profile and eliminate latency across CPU, GPU, memory, and I/O boundaries
Develop and optimize kernels for high-throughput, low-latency execution
Tune memory access patterns (global, shared, unified) for bandwidth efficiency
Optimize high-bandwidth sensor data ingestion
Identify development priorities by directly analyzing technical and physical system limitations in the field
Author technical documentation and mentor junior engineers on full system optimization