Loading description…
Vacancy description
Medcruit
$ 65 - $ 75 per unit
भारत
$65–$75 per task
Remote | Contract
We are seeking experienced forensic chemistry, toxicology, and chemical safety professionals to help evaluate how advanced AI systems handle technically complex and potentially dual-use questions.
The work focuses on determining whether AI models can distinguish between legitimate professional requests and requests that present genuine misuse or safety concerns. You will use your real-world casework expertise to create challenging evaluation scenarios, assess AI responses, and establish technically sound reference answers.
Develop challenging single-turn prompts based on your area of professional expertise.
Classify prompts across three categories: benign, dual-use, and adversarial .
Evaluate AI-generated responses against defined safety and policy standards.
Determine whether the model appropriately answered, limited, or refused a request.
Write reference answers explaining what a correct response should look like.
Provide clear technical reasoning supporting each evaluation.
Create scenarios that accurately reflect the boundary between legitimate professional casework and potentially harmful requests.
We are particularly interested in professionals with hands-on casework experience in areas such as:
Forensic chemistry involving controlled-substance identification and quantitation.
Clandestine laboratory response , precursor identification, and scene chemistry.
Forensic toxicology , including postmortem or human-performance casework.
Seized-drug and precursor trend monitoring or intelligence analysis.
Chemical casework involving expert reports or testimony .
Chemical safety and related forensic or analytical chemistry disciplines.
Prior experience with AI red-teaming or model evaluation is helpful, but the primary requirement is strong domain expertise and professional judgment.
This is highly writing-intensive and analytical work . Each evaluation requires a clear rationale that explains the technical judgment in a way that can be understood by non-specialists.
Strong candidates will demonstrate:
Excellent technical and scientific writing.
Ability to distinguish legitimate professional questions from potentially harmful requests.
Strong understanding of forensic chemistry, toxicology, or chemical safety.
Experience interpreting technical evidence and defending professional conclusions.
Ability to explain complex chemical concepts clearly and precisely.
Strong attention to detail and consistency when applying evaluation standards.
Published research, technical writing, expert witness experience, or comparable professional documentation experience is a strong advantage. Candidates may include a relevant writing sample or publication where appropriate.
The work must not involve or rely upon:
Classified information.
Export-controlled information.
Confidential information protected by an NDA.
Prepublication material subject to review or disclosure restrictions.
Professionals who have these obligations may still be considered; the project scope can be structured to avoid restricted material.
Remote contract engagement.
Work is focused on specialized expert evaluation and written technical assessment.
You may pause or step away from the work when needed, particularly given the nature of reviewing misuse-related scenarios.
The work is designed for experienced practitioners whose professional judgment can distinguish legitimate technical use from potentially dangerous applications.
Apply your forensic chemistry, toxicology, or chemical safety expertise to advanced AI evaluation.
Help assess whether AI systems handle sensitive technical questions appropriately.
Contribute professional judgment that general-purpose evaluation cannot reliably provide.
Work on specialized, technically challenging AI safety and model-evaluation projects.
Apply to bring your professional chemistry expertise to the evaluation of advanced AI systems.