Problem mapping
Before writing any code, I work out what the system really has to do: the boundaries, the constraints, the failure cases and what good enough looks like. It saves a lot of rework later.
Services
Where I can help.
I can help at any stage, from figuring out what the actual problem is to keeping an eye on robots that are already in the field. Most projects cover one or two of the steps below.
Architecture
From there I design the software architecture: components, interfaces and data flow that fit your stack and hardware. I write down the trade-offs so your team can build on it without me.
Implementation & quality
ROS nodes, planning and perception code in C++, Python or Rust. CI, tests and code standards are part of the setup from the start, not something to add later.
Simulation & testing
Simulation environments in IsaacSim, Mujoco or Gazebo, plus scenario tests, so behavior is proven and regressions are caught before you go back to the hardware.
Integration issues
When simulation, sensors and robot behavior disagree, I isolate the failure. Usually it comes down to timing, interfaces or coordinate frames. I fix it and add checks so it stays fixed.
Observability
Logging, metrics and dashboards for deployed robots. When something goes wrong in the field, you can see what actually happened instead of guessing from a bug report.
Not sure if I fit?
Send a short note on the platform, stack and the problem. I will tell you quickly whether I am the right person.
Get in touch