We treat every project as a small research program.
That means a clearly stated question, a hypothesis worth testing, a written record of what we tried, and an artifact at the end that you can keep using.
Areas we like
Physical systems
Sensing, control, and instrumentation for processes that have to actually work in the world.
Applied probability
Forecasting under uncertainty. Decision systems that quantify what they do not know.
Industrial software
Systems for engineers and operators. Tools that survive contact with reality.
Machine learning
Models scoped to specific decisions, with honest evaluation. No demoware.
What we believe
A working prototype is the strongest argument.
We would rather show than tell. The fastest path to clarity is usually a thing you can run.
Write everything down.
Memos outlive meetings. We treat clear writing as part of the deliverable, not overhead.
Generalists ship.
The best research engineers we know are generalists who go deep when it matters. We hire for taste and range.
Honesty over polish.
We will tell you when an approach will not work, when a deadline is unrealistic, and when you should hire someone else.