We build autonomous interceptor drones — fast, precise, and field-ready. Our intercept stack runs on embedded hardware at the edge: low-latency compute, real-time sensor pipelines, and firmware that has to work the first time, every time.
This role sits at the intersection of performance engineering and systems hardware. You'll own the compute layer that everything else depends on.
What you'll work on
As our Embedded Systems Engineer, you'll own the edge compute stack — maximizing what we can extract from the hardware that our seeker and intercept systems run on. You'll work closely with our GNC and CV engineers, who own the algorithms; your job is to make those algorithms fast, reliable, and deployable on constrained hardware.
Day-to-day, most of the work is applied software engineering: integration, optimization, tooling, and debugging. But you'll also shape how we scale beyond our current hardware.
Your responsibilities include:
Own and optimize our embedded compute stack on NVIDIA Jetson-class hardware — maximize GPU utilization, minimize latency, squeeze every cycle
Tune inference and video pipelines for edge deployment: CUDA kernel optimization, TensorRT, NVDEC, and memory layout for the Jetson architecture
Manage the full embedded software environment: BSP, kernel configuration, device drivers, power profiles, and boot optimization
Build tooling and CI/CD infrastructure for cross-compilation, OTA updates, and remote fleet management
Integrate peripherals and payloads: cameras, IMUs, networking interfaces, and custom hardware
Debug performance bottlenecks end-to-end — from hardware bring-up to application-level profiling
Collaborate with CV and GNC engineers to co-design pipelines that meet real-time constraints
Over time: contribute to the architecture and eventual design or outsourcing of a custom carrier board or SoM optimized for our seeker module
our stack
We're explicit about this so there's no guessing:
Edge hardware: NVIDIA Jetson (Orin / AGX class)
OS: Linux (Ubuntu-based, Jetson Linux BSP)
Compute: CUDA, TensorRT, NVDEC
Autonomy: Python on companion compute; C++ for performance-critical paths
Autopilot: ArduPilot (custom fork)
Comms: MAVLink / MAVSDK, LTE
No ROS
You're not expected to have worked with this exact combination before — but you should be able to ramp quickly and bring strong mental models to unfamiliar tools.
What We’re Looking For
Core Skills
Deep hands-on experience with NVIDIA Jetson or comparable embedded GPU platforms (Xavier, Orin, AGX)
Strong proficiency in CUDA and TensorRT — you know how to profile, identify bottlenecks, and write performant kernels
Solid Linux systems experience: kernel configuration, device trees, BSPs, systemd, and driver integration
C++ proficiency in embedded or systems contexts
Experience deploying and optimizing real-time inference pipelines at the edge
Comfortable with hardware bring-up: connecting, configuring, and debugging real peripherals
Methodical debugging mindset — you trace problems from silicon to software
Strong written and spoken English
Nice-to-Haves
Experience with video decode pipelines (NVDEC, V4L2, GStreamer) — particularly relevant to our seeker architecture
Familiarity with MAVLink, ArduPilot, or autopilot ecosystems
PCB design or hardware systems experience — even at the schematic review level
Experience specifying or working with custom carrier boards or system-on-modules
Background in defense, aerospace, or other safety-critical embedded environments
Python fluency (our autonomy layer is Python-heavy)
Docker and cross-compilation experience
Gazebo simulation
A1–A3 or A2 drone license
(Not required, but supported and considered a perk of the role)
Why join us?
You're shipping to systems that fly autonomous intercept missions. The hardware matters, the latency matters, and the margin for error is narrow.
Architecture ownership — You'll define how our compute layer evolves. That includes a longer-term path toward custom hardware purpose-built for our seeker platform — a rare opportunity to shape silicon-level architecture at a startup.
Equity that compounds — 2% structured via STAK, vesting over three years. At current valuation that's roughly €200k in equity — on top of a competitive base. As the company scales, so does the stake.
A mission worth working on — Europe needs sovereign autonomous defence capability. We're building it in Amsterdam in our R&D lab and we're moving fast.
A collaborative, hands-on culture — Trusk is a startup with a small, highly motivated core team that has skin in the game.
The application process
Three short rounds, including a practical technical assessment respectful of your time. We move quickly — typically two to four weeks from application to offer.
Trusk Technology
What's the interview process exactly?
When do I start?
Do you sponsor visas?
I've never worked in a startup, is that a problem?
Can I work fully remote?
Do you only focus on Ukraine?
Background check, what's that all about?
Still have more questions? Email us