Permit Review Expert (Part-Time Contractor)
About Brain Co. Brain Co. is at the leading edge of productionizing artificial intelligence for real world problems, propelling advancements and shaping the future with pioneering work. Our mission is to ensure that the transformative powers and automation of AI are applied to real world problems in various governments where manual work is still dominant. About the Role Brain Co. is building the world's first AI-native permitting, and we need people who've actually done the work to help us get it right. You would play a role in helping shape AI Permitting in the US. We're looking for former permit reviewers, plans examiners, and building officials to join us as part-time advisors. You'll work directly with our product and GTM team, flagging what our AI gets wrong, telling us what a real reviewer would do differently, and helping us build something that actually works in practice. We are looking for someone for a minimum of 2 hours per week for 3-4 months. Opportunity to extend as needed. What you'll do Interpret the codes - tell us what actually matters in your jurisdiction, what the edge cases are, where you routinely get tripped up Use the product - run real applications through it alongside your normal workflow Tell us where we are wrong - Review AI output on real permit submissions and flag errors, gaps, or misapplied code Jurisdiction knowledge - Help us understand how review actually works across different jurisdictions and permit types Define the benchmark - identify what is a pass / fail / conditional call that becomes the ground truth the model is measured against Who we're looking for 3+ years of hands-on review experience in Florida, Texas, or California, you know the local amendments, the common mistakes, and which submittals are going to come back before you've finished reading them Active or former plans examiner, building official, or permit reviewer, municipal department or third-party firm, both welcome Enough volume behind you to tell us what a clean submission looks like, what reliably triggers a revision request, and what gets rejected outright Opinions about where the process breaks down and no hesitation sharing them, "this is wrong because..." is exactly what we need Open to sharing examples on applications (passed and failed) No technical background required, your job is to tell us if the AI gets it right, not to explain how it works