We are a small group of engineers who like electronics, software, and the simulation that connects the two.
The team comes from electronics, embedded software, and simulation. We are as comfortable with a soldering iron and an oscilloscope as with a SPICE netlist or a build pipeline, and we move between hardware and software instead of treating them as separate problems.
Most of our work sits where analog and digital electronics, firmware, and data meet. We model and simulate before we build, automate the repetitive parts of testing, and keep everything version-controlled so a result can be reproduced months later.
We keep the team small by choice — it is the only way we know to stay close to the details. We take on work we can do well, explain it in plain terms, and hand it over so it can be maintained without us.
If a behaviour can't be written as an equation, a model, or a bounded approximation, we don't understand it yet — and we won't build on it.
"Faster", "lower-noise", "robust" mean nothing without a figure, a unit, and the exact conditions it was measured under.
No board is laid out until the SPICE result and the spec agree. The model is the contract — not the optimism.
Bench data overrules any opinion in the room, ours included. If the scope disagrees with the theory, the theory loses.
A result that works once, on one rig, on one afternoon is a story. It has to rerun from a clean checkout to count.
A model can suggest a design; only the simulator and the bench can accept it. We never ship what we can't independently verify.
~9.4M
60+
25+
8×
AI earns its place only when it touches something measurable. We wire language models into the tools we already trust — SPICE, the netlist, datasheets, the test rig — and check every suggestion against simulation or bench data before it reaches a board. Not a demo; shorter design loops on real hardware.