AI Tutor, Civil Engineering Specialist (contract), Handshake AI

📋 Description

• • Design and evaluate high-difficulty Civil Engineering prompts that challenge the reasoning limits of large language models (LLMs) in civil engineering domains.
• • Identify reasoning errors, logical gaps, and failure modes in AI-generated responses related to structural mechanics, geotechnical engineering, transportation engineering, and fluid dynamics.
• • Apply adversarial prompting techniques to expose weaknesses in model understanding and push AI systems beyond surface-level accuracy.
• • Provide expert-level, detailed feedback on AI model outputs to correct technical inaccuracies, incorrect formulas, flawed assumptions, and misapplied design standards.
• • Ensure technical precision in numerical calculations, code-based solutions, and regulatory references (e.g., AASHTO, ACI, FEMA) within model responses.
• • Collaborate with project leads to establish and maintain quality benchmarks for Civil Engineering content across AI training datasets.
• • Contribute to the refinement of evaluation rubrics used to score AI responses for correctness, completeness, and pedagogical clarity.
• • Review and validate model outputs against established engineering principles, industry codes, and peer-reviewed literature.
• • Flag inconsistencies between model-generated solutions and real-world engineering practice, including oversimplifications or missing boundary conditions.
• • Document specific failure patterns observed in AI responses to inform iterative improvements in model training and alignment.
• • Maintain strict attention to detail in technical communication, ensuring all feedback is clear, actionable, and free of ambiguity.
• • Work independently to assess complex engineering problems presented in open-ended prompts, with no reliance on standard answer keys.
• • Contribute to the creation of novel problem scenarios that test advanced application of civil engineering concepts under constraints.
• • Participate in periodic alignment sessions to ensure consistency in evaluation standards among other specialist reviewers.
• • Adhere to data quality protocols that prioritize accuracy over volume, prioritizing depth of analysis over quantity of inputs.
• • Ensure all feedback is grounded in accepted civil engineering practices and does not introduce speculative or non-standard interpretations.
• • Operate in a fully remote, asynchronous environment with flexible weekly hours (capped at 40 hours/week; no minimum requirement).
• • Maintain confidentiality of proprietary AI datasets, prompts, and evaluation methodologies provided by Handshake Technologies, Inc.
• • Support the broader mission of improving AI reliability in STEM education and professional engineering contexts through rigorous data curation.