Stephan Claxton
Most high-consequence software is built without a system. I work on the engineering discipline that fixes that, across autonomy, space, and defense.
Today: systems engineering for autonomy at Applied Intuition. Before: MBSE for space and defense programs at Lockheed Martin, then technical program management at Strategic Technology Consulting (Arcfield).
Field notes
- Structure produces behavior. The architecture of a system is its actual design. Change the architecture and you change what the system does. Change anything else, and you mostly change the story you tell about it.
- Building autonomy is the hardest verification problem in engineering today, and the one rewriting how systems engineering is practiced.
- Requirements traceability is institutional memory. It feels useful exactly once, on the worst day. That is enough.
- The aerospace tradition gave systems engineering its methods. AI and autonomy are giving it back its relevance, and making executable models the new default.
Writing
- Doing systems engineering at startup cadence
The methods that work at Lockheed do not survive contact with a weekly release schedule. The methods that survive are the ones worth keeping anyway.
- The curse of the sheep
Reflections on context, mastery, and systems engineering from INCOSE IS 2026, and why the discipline exists to preserve understanding across time.
- Requirements traceability is a survival skill, not a compliance checkbox
What aerospace gets right about traceability, what autonomous vehicle development is still figuring out, and why the answer changes depending on who is asking the question.
- Your system model belongs in CI, not your document repository
SysML v2 helps, but the harder shift is how systems engineers work when AI handles the typing and git owns the truth.
- Reading Donella Meadows from inside an autonomy company
A book about ecosystems and global resource flows is the most useful thing I have read for autonomy engineering. Why Meadows's leverage points matter more for AV development than additional simulation miles ever will.
- What satellite systems engineering taught me about building autonomous vehicles
Two domains that look nothing alike on the surface, and the MBSE discipline that translates between them.