Note: The job is a remote job and is open to candidates in USA. Agency is seeking experts and researchers for an AI evaluation project focused on Product Deep Research. The main responsibilities include designing complex prompts, creating rubrics for evaluation, auditing multidimensional analyses, and verifying the factual accuracy of the AI model's research capabilities.
Responsibilities
- Designing and writing complex, multidimensional prompts that compel the AI model to execute deep, nuanced product analysis rather than surface-level summaries
- Generating auditable, and atomic rubric items per task. These rubrics must objectively evaluate the model’s response across mandatory analytical dimensions
- Evaluate the model's generated outputs to ensure it accurately weighs product trade-offs, evaluates competitive market positioning, and assesses long-term value
- Delivering feedback verifying the factual precision of the model's deep research capabilities across various product categories
Skills
- Demonstrable ability to source, verify, and synthesize complex product information from highly reliable sources
- The ability to break down complex consumer or enterprise products into critical evaluation vectors (e.g., build quality, ecosystem lock-in, price-to-performance ratio)
- Familiarity with structuring prompts to elicit highly specific, constrained, and analytical outputs from Large Language Models (LLMs)
- Broad familiarity with consumer or enterprise product landscapes and how to properly assess them
Benefits
- Health insurance
- PTO
Company Overview
- Invisible Technologies makes AI work. Our platform cleans, labels, and structures company data so it’s ready for AI. It was founded in 2015, and is headquartered in San Francisco, California, USA, with a workforce of 201-500 employees. Its website is https://www.invisible.co/.
Company H1B Sponsorship
- Invisible Technologies has a track record of offering H1B sponsorships, with 9 in 2025. Please note that this does not guarantee sponsorship for this specific role.