Applied Physics Specialist

About The Role

What if your deep expertise in physics could directly shape how AI understands the physical world — ensuring it respects the fundamental laws that govern everything from subatomic particles to cosmological systems? We're looking for PhD-level Applied Physicists to join Alignerr and help train cutting-edge AI models. You'll design research-grade physics problems, expose flaws in AI reasoning, and build the benchmarks that push these models toward genuine physical understanding. This is a fully remote, flexible contract role built for researchers and academics who want to contribute to something meaningful — on their own schedule.

Who You Are

• Holds a PhD (completed or near completion) in Applied Physics, Physics, Engineering Physics, or a closely related field
• Deep mastery of core physics pillars: Classical Mechanics, Electrodynamics, Statistical Mechanics, and Quantum Mechanics
• Exceptional ability to communicate complex physical phenomena and mathematical derivations in clear, structured written English
• Uncompromising attention to detail — units, notation, and logical flow matter to you
• Self-directed and comfortable working independently on technical tasks
• No prior AI experience required — your physics expertise is what counts
• Nice to have: Experience with data annotation, scientific dataset evaluation, or quality assurance; Proficiency with tools such as Python (NumPy/SciPy), MATLAB, or COMSOL; Background in academic research, technical writing, or curriculum development; Familiarity with AI or LLM evaluation workflows

What You'll Do

• Design Advanced Physics Problems — Craft open-ended, multi-step problems at PhD qualifying exam level, spanning quantum mechanics, electromagnetism, thermodynamics, and beyond
• Author Ground-Truth Solutions — Develop rigorous, step-by-step "golden responses" with precise physical constants, unit conversions, and airtight mathematical derivations
• Audit AI Reasoning — Evaluate AI-generated proofs and simulations for physical consistency, identifying where models "hallucinate" results that violate first principles
• Refined Model Behavior — Provide structured, expert feedback to improve AI reasoning around boundary conditions, conservation laws, and physics-informed constraints
• Benchmark AI Understanding — Contribute to university- and research-level evaluation datasets that test the limits of what AI truly understands about the physical world