PM for AI Specifications — Shape Web App Quality

Obsidian; New York (NY); On-site; USD 90,000 - 130,000; Full time

Job summary

Obsidian is seeking experienced Product Managers, Product Designers, and Technical Product professionals to review AI-generated web application specifications before they're used in model evaluation. You will assess realism, completeness, and clarity, ensuring AI-generated applications can be built and tested against high-quality specifications.

You will evaluate whether product requests are realistic, complete, and unambiguous, verify alignment with test scenarios, and provide structured

Qualifications

• 3+ years of experience in Product Management, Technical Product Management, Product Design, or Business Analysis.
• Experience writing or reviewing product requirements, PRDs, user stories, or acceptance criteria.
• Strong understanding of modern web application workflows and user experiences.
• Excellent written communication and attention to detail.

Responsibilities

• Review AI-generated product specifications for web applications.
• Evaluate whether product requests are realistic, complete, and unambiguous.
• Verify that subsequent feature requests build correctly on the existing product state.
• Ensure product requirements align with associated browser-based test scenarios.
• Identify inconsistencies, missing requirements, or unclear acceptance criteria.
• Provide structured written feedback using project review guidelines.

Skills

Product Management

Product Design

Technical Product Management

Business Analysis

Written communication

Tools

PRDs

User stories

Acceptance criteria

Job description

Obsidian is seeking experienced Product Managers, Product Designers, and Technical Product professionals to review AI-generated web application specifications before they're used in model evaluation. You will assess realism, completeness, and clarity, ensuring AI-generated applications can be built and tested against high-quality specifications.

You will evaluate whether product requests are realistic, complete, and unambiguous, verify alignment with test scenarios, and provide structured