AI Research Peer Review Evaluator (ML/AI)

Join; Zürich; Vor Ort; CHF 83.000 - 138.000; Teilzeit

Benefits dieser Stelle

Fully remote

Part-time contract

Zusammenfassung

Lightly AG, a Zurich-based AI company, seeks researchers with strong ML/AI backgrounds to support an AI evaluation project focusing on scientific peer review. You will evaluate agentic AI reviews against expert human peer reviews, using a structured rubric and evidence-based rationales.

This is a remote, project-based contractor opportunity with flexible hours. Part-time arrangement preferred for ongoing evaluation across multiple papers.

Qualifikationen

• Advanced analytical thinking and ability to interpret complex ML/AI content.
• Strong written communication skills for clear evaluation rationales.
• Excellent attention to detail when identifying factual inaccuracies or hallucinations.

Aufgaben

• Read and scan ML/AI research papers to extract core contributions and experiments.
• Review original human peer reviews to establish baselines for each paper.
• Evaluate AI-generated peer reviews against the expert baseline using a structured rubric.
• Assess technical accuracy, depth, constructive value, and novelty in each AI review.
• Identify hallucinations, unsupported claims, and missed technical issues.
• Compare two AI reviews to determine which provides stronger analysis.
• Search and verify relevant literature using Google Scholar, arXiv, or Semantic Scholar.
• Provide concise, evidence-based rationales applying the project rubric consistently.

Kenntnisse

Analytical thinking

Written communication

Attention to detail

Ausbildung

Master’s degree or PhD in ML/AI/CS

Jobbeschreibung

Lightly AG is a Zurich-based AI company and ETH/HSG spin-off, backed by Y Combinator and top-tier investors. Our machine learning and computer vision technology is trusted by global leaders in autonomous driving, medical imaging, and visual inspection.

We’re looking for researchers with strong Machine Learning / AI backgrounds to support an AI evaluation project focused on scientific peer review. You’ll evaluate reviews generated by agentic AI systems and compare them against expert human peer reviews of ML/AI research papers.

This is a remote, project-based contractor opportunity with flexible working hours.

Tasks

What you'll be doing

• Read and scan ML/AI research papers to understand their core contributions, methodology, experiments, and claims
• Review the original human peer reviews to establish an expert baseline for each paper
• Evaluate AI-generated peer reviews against that baseline using a structured scoring rubric
• Assess the technical accuracy, analytical depth, constructive value, and novelty/significance assessment of each AI review
• Identify hallucinations, unsupported claims, missed technical issues, or valuable insights surfaced by the AI reviewers
• Compare two AI-generated reviews side-by-side and determine where one provides stronger or more useful analysis
• Search and verify relevant academic literature using sources such as Google Scholar, arXiv, or Semantic Scholar, including checking whether cited prior work was available before the paper’s submission date
• Provide concise, evidence-based rationales explaining your evaluation decisions and consistently apply the project rubric

The evaluation specifically looks at whether agentic AI reviewers can provide meaningful value beyond expert human reviewers—for example, by identifying relevant prior literature that humans missed, questioning important assumptions, or resolving inconsistencies using evidence.

Requirements

You're a strong candidate if you:

• Have a Master’s, PhD, or are currently pursuing graduate study in Machine Learning, Artificial Intelligence, Computer Science, Statistics, or a closely related technical field
• Have contributed to at least one scientific/research paper, ideally as a first author, although co-authors and other substantial contributors are also welcome
• Have experience critically reading ML/AI research papers, including evaluating methodology, experimental design, results, limitations, and scientific claims
• Are familiar with major ML/AI research venues, such as NeurIPS, ICML, ICLR, ACL, CVPR, or comparable conferences and journals
• Have prior academic peer-review experience, ideally for an ML/AI conference or journal — strongly preferred
• Are comfortable conducting academic literature searches and verifying prior work, publication dates, citations, and novelty claims
• Have strong analytical and written communication skills and can distinguish meaningful technical concerns from superficial criticism
• Can provide clear, concise, evidence-based rationales for your decisions
• Can consistently apply detailed evaluation guidelines and scoring rubrics across multiple papers and reviews
• Have strong attention to detail, particularly when identifying factual inaccuracies or hallucinated technical claims

Benefits

• Fully remote and flexible — work from anywhere
• Part-time contractor role with flexible hours
• Work directly on the evaluation of cutting-edge agentic AI systems for scientific research
• Apply your ML/AI research expertise to help measure and improve the quality of AI-generated scientific peer review