Help make engineering judgment reusable.
Build the next generation of engineering intelligence on a working foundation: model-guided circuit search, learned layout guidance and experience-backed planning. Work directly with the founder to turn these capabilities into a shared workspace where engineers and AI explore, test and improve physical designs together.
What you would own
Extend a working foundation: probabilistic circuit models, learned layout guidance, an experience store for block search and system planning, and simulator-backed verification. Own how the system learns from experience, chooses experiments and brings its reasoning into the engineer's workflow. Build the models, evaluation and interactive software that take the capability to harder problems.
What you bring
Probabilistic machine learning and Bayesian optimisation or active learning, applied. You have put a model in front of a solver, a simulator or an instrument before, and you know when to use an existing model, classical optimisation or new research. Circuit knowledge is welcome, not required; the founder supplies the domain.
We are starting conversations now for a founding role, with a start date dependent on funding.