Robotics Integration Engineer - Simulation and Field Validation

Type: Full-time, permanent employment Location: Madrid, on-site Relocation: Relocation support and visa sponsorship are available Salary: €60,000–€75,000 gross base salary per year, depending on experience and demonstrated depth, plus equity

About Nonlimi
Nonlimi builds simulation and validation infrastructure for autonomous systems. We want robotics teams to connect the autonomy stacks they already run, exercise them against reproducible scenarios, and carry evidence from field failures back into simulation without rebuilding their systems around us.

Our near-term users may build warehouse and industrial robots or civilian drones. Their interfaces differ — ROS2 and machine telemetry on one side, PX4/MAVLink and control-system models on the other — but the product loop is the same:

external autonomy stack → simulation → evidence → field log → replay → diagnosis and tuning

You will own that loop.

The Role
You will make Nonlimi work with autonomy stacks that were not designed for it, and own the diagnosis when the first attempt fails. Boundary conditions, timing assumptions, build compatibility, protocol drift and mismatched models are where integrations actually break.

This is a hands-on software and systems role. You will design interfaces, write or extend adapters, inspect runtime behaviour and turn field evidence into reproducible tests. The work begins with our ROS2 surface and field-evidence loop, then carries the same product pattern into civilian-drone stacks as priorities and evidence develop.

You will work directly with the people responsible for Nonlimi's deterministic simulation core and evaluation runtime. Your job is to make external systems usable with that product without turning each integration into a bespoke project.

What You Will Do
Design, implement and version the interfaces between Nonlimi and external autonomy stacks, including middleware graphs, flight controllers, control-system models and their data contracts.
Build or extend the adapters, bridges and wrappers needed for closed-loop operation; this includes systems-level code, not only configuration.
Diagnose failures across software, middleware, networking, timing and model boundaries, then isolate which side of the interface is responsible.
Turn logs and telemetry from real or representative systems into reproducible replay, comparison and tuning workflows.
Build automated integration tests, regression fixtures, compatibility records and a replayable failure corpus so each integration improves the product.
Work with the core and evaluation-runtime owners to define clear interface contracts and readiness criteria.
What Success Looks Like
External robotics and autonomy stacks can connect to Nonlimi through clear, reliable interfaces without being rebuilt around the platform.
Integration failures can be isolated across the runtime boundary and resolved through durable product or adapter improvements.
Field logs and telemetry can move through a reproducible replay, diagnosis and comparison workflow.
Each integration leaves behind reusable compatibility knowledge, failure cases and regression evidence that make the next integration easier.
The same validation product can support different robotics domains through focused adapters rather than separate one-off projects.
What You Bring
Core evidence
You have made an external autonomy stack work against a simulation or validation platform it was not designed for, and you personally diagnosed and resolved why the integration broke.

Equivalent evidence against any comparable simulation or robotics platform is the bar. Proving it against Nonlimi becomes the first assignment, not a prerequisite.

Required
Substantial production experience with robotics, autonomy or another distributed real-time system.
Ownership of a robotics runtime or middleware boundary such as ROS2/DDS, PX4/MAVLink, a proprietary autonomy stack or a comparable control-system interface.
Ability to read, write and debug systems-level integration code in C++ on Linux.
Demonstrated depth in at least one of these areas: PX4/MAVLink SITL, Simulink/MATLAB co-simulation, or production field-log replay and tuning.
Clear written and verbal communication across software, controls and product boundaries.
ROS2 experience is valuable, but this role also requires depth in at least one of the three areas above.

Strongly preferred
Production ROS2 depth in a system that shipped and ran.
Depth in a second surface from the list above; experience spanning both ground robotics and civilian drones is particularly useful.
Experience with software-in-the-loop or hardware-in-the-loop validation, distributed-system timing, networking or time synchronization.
Python or shell experience for diagnostic tooling and test automation.
Experience integrating a third-party simulation platform, autonomy stack or robotics API.
Experience building compatibility matrices, diagnostic tooling or regression fixtures from repeated field failures.
Not required
Deep aerodynamics or blade-element rotor modelling.
Perception-model training or graphics specialization.
Sales or pre-sales ownership.
Role Boundaries
You will collaborate closely with the owners of simulation-core reliability, perception fidelity and product packaging, but those are not this role's primary remit. We are looking for an integration owner, not a specialist in aerodynamics, perception-model training, cloud infrastructure or commerce.

Who Tends to Fit
This trait often appears in robotics platform engineers, field robotics engineers, flight-controller integration engineers, and control-systems engineers who also write production software. If your work repeatedly placed you between a simulator, a controller and a real machine — and you became the person who could explain which assumption was wrong — that is the shape we are looking for.

What We Offer
€60,000–€75,000 gross base salary per year plus equity, calibrated to experience and demonstrated depth.
Relocation support and visa sponsorship for joining the team in Madrid.
Direct product ownership of the interoperability and field-evidence loop, with room to improve the underlying product rather than repeatedly work around it.
Close collaboration with the core team responsible for deterministic simulation and evaluation-runtime behaviour.
Cross-domain impact: build one validation product that can support warehouse, industrial and civilian-drone systems through reusable interfaces and evidence.