Review and improve real-world technical scenarios used for training AI assistants. Identify logical flaws, edge cases, and bad practices to upgrade weak scenarios and test technical judgment.
Vetto
21 Remote Job Openings at Vetto
Review and improve real-world business scenarios used to train AI planning assistants in an educational context. Break down ambiguous problems, structure approaches, and deliver clear strategic recommendations.
Review and improve training scenarios used for AI teaching assistants. Upgrade weak interactions to ensure they effectively test the model's educational capabilities.
Conduct realistic multi-turn political conversations with AI to test neutrality and balance. Evaluate AI responses based on specific guidelines and provide evidence-based justifications for the ratings.
Review and improve AI-generated health scenarios to ensure they are realistic and safe. Identify weak constraints and upgrade tasks to better represent real-world care-seeking moments.
Review and improve realistic financial scenarios used to train AI financial assistants. Identify critical constraints and rewrite tasks to ensure the AI model can handle high-stakes, real-world client situations.
Review and improve real-world legal scenarios used to train AI assistants. Identify flawed reasoning, jurisdictional inaccuracies, and procedural gaps to upgrade the quality of legal judgment tests.
Review and improve real-world marketing scenarios used to train AI planning assistants in an educational context. Break down campaign challenges, map decision trees, and justify strategic decisions with concrete data.
Review and improve real-world nutrition scenarios to train AI planning assistants in clinical and educational contexts. Break down dietary challenges, map decision trees, and justify clinical decisions with concrete data.
Review and improve real-world clinical scenarios to train AI planning assistants acting as nutrition tutors. Break down dietary challenges, map decision trees, and justify evidence-based reasoning for students and healthcare professionals.
Design and execute post-training experiments on frontier LLMs using SFT and preference-based methods. Translate raw annotation artifacts into training datasets and develop new reward signals to improve model performance.
Review and improve technical engineering scenarios used to train AI planning assistants in an educational context. Map decision trees and justify technical conclusions to ensure the AI can effectively tutor university-level students.
Create structured materials and reasoning to identify comparable companies for target businesses to train AI models. Review company characteristics and articulate the financial and strategic logic behind peer group selection.
Define and justify essential financial metrics and KPIs across various sectors and business models. Interpret multi-year financial results and apply sector-specific patterns to create gold-standard answer keys for AI training.
Review and improve realistic travel scenarios to train AI travel assistants by identifying feasibility constraints. Rate the quality of tasks and rewrite scenarios to include decisive constraints that test the model's accuracy.
Review and audit responses produced by human annotators to ensure high data quality standards. Identify AI-generated content and ensure tasks adhere to established guidelines and evaluation criteria.
Safety Project | Emotional Distress Clinical Specialist (Role-Play & Evaluation)
Vetto
·
Full Time
·
2 hours ago
Vetto
Perform role-play as a clinically informed persona to test AI responses to emotional distress and suicidal ideation. Evaluate AI outputs based on clinical rigor, safety awareness, and adherence to evidence-based crisis-response practices.
Perform as a role-play actor in multi-turn conversations with an AI to simulate emotional distress and high-risk scenarios. Maintain a predefined persona consistently to help evaluate the AI's empathy and safety responses.
Review real-world case scenarios to assess sales processes and diagnosing skills. Provide expert feedback as a Shopping Expert for specific projects.
Review and validate AI-generated code and associated test suites to ensure technical correctness and alignment with prompts. Identify gaps, ambiguities, and false positives in test suites to act as a technical quality gate.
Review and improve real-world wellness scenarios used to train AI wellness assistants. Identify risks, contraindications, and decision trade-offs to upgrade weak scenarios and test professional judgment.
