Chemist AI Safety SME – Computational Chemistry

DeepMind Technologies Limited; Mountain View (CA); Sur place; USD 174 000 - 252 000; Plein temps

Résumé du poste

Google DeepMind’s CBRNE team invites qualified chemists to join as a Chemist Subject Matter Expert (SME) in the Responsible Frontier AI Research (RFAIR) group, bringing broad chemistry expertise to safety-focused AI evaluation.

The role focuses on evaluating and mitigating safety risks of frontier AI models (primarily LLMs) in the chemistry domain, advising on mitigations while preserving legitimate science.

Qualifications

• PhD in synthetic or computational chemistry, or equivalent practical experience.
• 2 years of post-doctoral, industry, or defense experience, including wet-lab bench experience designing, executing, and troubleshooting multi-step syntheses.
• Experience in cheminformatics and digital drug discovery tooling, specifically using Python (e.g., RDKit) to analyze reaction and bioactivity/toxicity datasets (e.g., QSAR), alongside ML (e.g., protein-ligand co-folding) or physics-based (e.g., FEP) binding affinity prediction.
• Experience with computer-aided synthesis planning (CASP), retrosynthesis tools, and 3D structure-based modeling, such as molecular docking and pharmacophore modeling.
• Experience working with Scientific AIs and applying LLMs to scientific problems.

Responsabilités

• Design, execute, and critically review the outputs of computational chemical evaluations, including automated synthesis planning, predictive toxicology, and molecular docking, to test the maximum agentic capabilities of our AI models.
• Review model outputs of synthetic chemistry evaluations, commenting on key metrics (such as accuracy) and the potential uplift provided by the model in the context of a given threat actor.
• Provide recommendations for implementing mitigations to ensure the chemical safety of our models, while minimizing the impact on legitimate science.
• Use information provided from evaluations and other data sources to update harm frameworks and threat models.
• Communicate technical results and commentary to audiences with a range of expertise, from fellow chemists to executive decision-makers.

Connaissances

Cheminformatics

Python (RDKit)

Synthesis planning

LLMs in science

Formation

PhD in synthetic/computational chemistry

Outils

RDKit

CASP tools

Molecular docking

3D structure modeling

Description du poste

Google DeepMind’s CBRNE team invites qualified chemists to join as a Chemist Subject Matter Expert (SME) in the Responsible Frontier AI Research (RFAIR) group, bringing broad chemistry expertise to safety-focused AI evaluation.

The role focuses on evaluating and mitigating safety risks of frontier AI models (primarily LLMs) in the chemistry domain, advising on mitigations while preserving legitimate science.