Synthetic / Process Chemist for AI Red-Teaming

Apply hands-on synthetic and process chemistry expertise to help evaluate how advanced AI systems handle technical chemistry requests, distinguishing legitimate scientific support from requests with meaningful misuse potential. Role Overview

This writing-intensive, per-task engagement focuses on testing AI responses at the boundary between routine professional chemistry and dual-use risk. You will use practical experience in route design, scale-up, process safety, impurity analysis, and regulatory chemistry to assess whether models provide appropriate, policy-aligned answers.

• Create challenging single-turn chemistry prompts classified as benign, dual-use, or adversarial.
• Evaluate AI responses against a defined policy standard and determine whether each response was handled correctly.
• Develop reference answers that show the appropriate response and explain the underlying technical reasoning.
• Write clear rationales that non-specialists can understand.

• Practical experience running and scaling synthetic routes from bench through pilot and plant environments.
• Relevant background in one or more of the following: synthetic route design and development in pharmaceutical, agrochemical, or fine chemical settings; process chemistry and reaction engineering, including optimization, kinetics, and impurity control; scale-up and associated safety review; precursor sourcing, controlled-substance registration, or regulatory chemistry; or method development and troubleshooting in GMP or an equivalent quality environment.
• Ability to distinguish ordinary process-chemistry questions from requests intended to elicit harmful or inappropriate assistance.
• Strong technical writing skills. Published research, prior technical writing, expert-witness experience, or red-teaming experience is valuable. Include a writing sample or link with your application.
• Comfort reading and writing about chemistry misuse scenarios for sustained periods. You may pause or step away at any time without penalty.

• Remote, per-task engagement.

• 65 to 75 per task.

• Do not use or rely on classified information, export-controlled information, NDA-covered material, or information subject to prepublication review obligations.
• Candidates with such obligations may still be considered if they disclose them in the application so the work can be scoped appropriately.