Obtenez une réponse de cet employeur
Votre CV et votre lettre de motivation, adaptés exactement à ce que l’employeur recherche.
Innodata Inc. is recruiting US-based experts to review AI-generated building-code determinations on architectural plans. You will verify, correct, and justify pass/fail outcomes with cited sources, contributing high-quality training signals for the system.
The role requires hands-on permitting experience (2+ years) and strong knowledge of IBC/IRC codes, with preference for those familiar with certain states. This is expert review work, not data entry, with competitive hourly pay.
Innodata (Nasdaq: INOD) is a global data engineering company. We believe that data and Artificial Intelligence (AI) are inextricably linked. Our mission is to enable the responsible advancement of artificial intelligence by providing the data, evaluation frameworks, and human expertise required to build AI systems that can be trusted at scale. We provide a range of transferable solutions, platforms, and services for Generative AI / AI builders and adopters. In every relationship, we honor our 36+ year legacy delivering the highest quality data and outstanding outcomes for our customers.
We are recruiting US-based architects and certified plans examiners to evaluate the output of an AI system that performs automated building-code compliance review on architectural permit applications.
The system ingests permit submissions — new builds, alterations, modifications — and issues a compliant or non-compliant determination for each applicable check, drawn from the International Building Code. Every determination comes with a cited code provision and a marked location on the drawing set showing the evidence relied upon.
Your job is to tell us where the system got it wrong. You are applying the same judgment you already apply in plan review or permit preparation, to a set of determinations that have already been made. Where the system is correct, you confirm it. Where it is wrong, you correct it and the correction becomes training signal.
This is expert review work, not data entry. There is no volume of repetitive clicking to grind through and no generalist labeling pool to compete with — the entire premise of the program is that these judgments cannot be made by anyone without real code expertise.
The expected hourly range for this position is $50-$75 p/hour, based on experience, skills, and qualifications.