Biosafety Environments Lead — AI Risk & Standards

Collinear AI, Inc.; United States; On-site; USD 100,000 - 140,000; Full time

Job summary

Collinear AI is seeking a domain authority to lead its biosafety track. You will translate everyday biosafety judgments into actionable tasks, reference solutions, and graded rubrics to measure model reasoning about biological risk like a trained expert. You will define when to hedge, decline, or answer fully, with a focus on safety and regulatory compliance.

No prior AI experience is required; we will teach tooling and workflow while you bring field expertise and judgment to the role.

Qualifications

• PhD in biology or related field, or a master’s with 4+ years biosafety experience.
• Hands-on lab experience with BSL-2 or higher.
• Familiarity with biosafety frameworks (BMBL, NIH Guidelines, OSHA).
• Ability to write and review risk assessments and IBC submissions.
• Experience with large language models and evaluation pipelines.

Responsibilities

• Map biosafety work into tasks, risk assessments, and containment decisions.
• Write expert prompts and gold-standard answers with peer-reviewed references.
• Build rubrics and verifiers to define complete and safe answers.
• Grade model outputs for accuracy, regulatory compliance, and risk.
• Red-team prompts to probe boundary between safe guidance and uplift.
• Shape environments with engineers for workflow state and rewards.
• Raise the bar by reviewing tasks and setting biosafety standards.

Skills

Biosafety risk assessment

Clear writing

Prompt engineering

Rubrics design

Model evaluation

LLM tooling

Education

PhD in biology or related field

Master’s +4 years biosafety experience

Tools

Python

Job description

About the role

We’re looking for a Member of Technical Staff, Research focused on drug discovery to help advance AI’s ability to reason about chemistry and solve problems that matter to discovery teams.

You’ll bring deep expertise in computational and medicinal chemistry, along with hands-on experience building and evaluating machine learning models. Your work will connect model development to real scientific challenges: designing compounds, planning syntheses, interpreting experimental results, and navigating tradeoffs across potency, selectivity, and ADMET properties.

This is a hands-on research role with broad ownership. You’ll develop datasets, evaluations, and reinforcement learning environments; run experiments with our AI researchers; and help decide which problems we should tackle next. As an early contributor to this effort, you’ll help define both the research agenda and how we work.

What you’ll do

• Own research from idea to result. Identify high-value drug discovery problems, define a research approach, build prototypes, and test whether they improve model capabilities.

• Build environments that teach AI chemistry. Develop reinforcement learning tasks with meaningful rewards, reliable grading, and safeguards against shortcuts and data leakage.

• Turn scientific data into useful training signals. Curate and structure data from literature, patents, experimental results, and other relevant sources.

• Measure progress on real discovery tasks. Build evaluations for molecular design, reaction prediction, synthesis planning, structure elucidation, and multiparameter optimization.

• Run experiments and follow the evidence. Work closely with AI researchers and engineers to train models, investigate failures, and turn findings into the next round of improvements.

• Keep the science grounded. Assess computational results against experimental evidence and collaborate with scientific partners to validate promising approaches.

• Build for a growing team. Create reproducible workflows, document key decisions, and communicate results clearly so others can build on your work.

What you’ll bring

• A PhD in chemistry, computational chemistry, cheminformatics, or a related field—or equivalent research experience.

• Deep knowledge of organic and medicinal chemistry, including reactions, synthesis, structure–activity relationships, and the practical tradeoffs involved in drug discovery.

• Strong Python skills and a solid foundation in machine learning, with hands-on experience training and evaluating models.

• Scientific judgment: you can distinguish a promising benchmark result from a meaningful advance in a discovery workflow.

• Comfort working through ambiguity, choosing priorities, and taking a project from an open question to a concrete result.

• A collaborative approach, clear communication, and the willingness to work across research and engineering as the team grows.
  

Experience that would stand out

• Reinforcement learning, large language models, scientific agents, or chemistry-focused evaluations and benchmarks.

• Contributions to drug discovery programs in close collaboration with medicinal or synthetic chemists.

• Experience curating reaction, bioactivity, ADMET, or spectroscopic datasets, with attention to provenance, quality, and data leakage.

• Familiarity with tools and methods such as RDKit, molecular docking, molecular dynamics, free energy calculations, or quantum chemistry.

• Hands-on synthetic or medicinal chemistry experience at the bench.

• Publications, patents, open-source tools, or deployed research systems that demonstrate scientific impact.

You don’t need to check every box. We’re interested in people who bring depth, curiosity, and the drive to build useful things—and who want to help shape what comes next.