Founding Engineer — Physics × EE × ML

Why now

AI has changed how software gets written. Physical engineering is next: the chips, boards, radios and power systems that everything else runs on are still designed with decades-old tools, by intuition, a few iterations at a time. Models that can reason over a design, run the physics and check their own work will change that, and the stack that makes it possible is being built now.

About Faraday Loop

We build AI for physical engineering. We work at the point where learned models, physics simulation and real engineering design meet, with real design teams, on problems where the answer has to hold up against physics. We are backed by Founders, Inc. and are part of NVIDIA Inception.

About the role

You'd be one of the first engineers, and you'd shape what we build as much as how. We need physics, electrical engineering and machine learning in some combination. Nobody has all three at depth; we expect real depth in one, working fluency in a second, and curiosity about the third.

What you'll do

• Build and validate simulation and modelling systems, and check them against references and measurement.
• Train models on physics, and design the evaluations that show where they can and can't be trusted.
• Build agents that do real engineering work and verify their own results.
• Work directly with engineers at design partners on their real problems.
• Own a problem end to end, from the physics to the thing a customer runs.

You should have

• Deep, hands-on experience in at least one of: computational physics or numerical simulation; electrical engineering design (RF, signal or power integrity, analog, packaging); machine learning for physical systems.
• Working fluency in a second of those.
• A track record of taking something from idea to a working result that someone else used.
• Strong Python, and the habit of checking a result against a second method before believing it.
• Comfort with an early company: ambiguous problems, few people, direct customer contact.

Nice to have

• You've written, contributed to or validated a solver.
• Experience training models on simulation data.
• HPC or GPU experience, or numerical linear algebra.
• A PhD or equivalent research experience, publications, or open-source scientific software.
• Experience building LLM agents with tools and verification.

How to apply

Apply here with your CV or LinkedIn, and a link to or description of one thing you built. Tell us what was hard about it. If you don't match every line above, apply anyway; we would rather read about the thing you built.