Software Engineer, AI Safety and Biosecurity

your Jared; Northern (KY); Hybrid; USD 70,000 - 115,000; Full time

Benefits offered by this job

Remote work

Job summary

SecureBio seeks a Software Engineer to scale and systematize evaluations of frontier AI models, focusing on biosecurity and misuse risk. You’ll own running the evaluation suite via Inspect, maintain dashboards, and provide structured outputs for reports.

This role advances our mission to reduce catastrophic biological risks by ensuring robust, reproducible, timely assessments. You’ll work remotely with scientists and engineers, contribute to code that supports pre-release testing, and help shape

Qualifications

• 1-2 years of professional software engineering experience, particularly with Python.
• Familiarity with basic statistics, data analysis, and data visualization.
• Proficiency in using large language models in personal or professional contexts.
• Demonstrated ability to navigate and understand unfamiliar codebases.
• Strong communication skills, responsiveness, attention to detail, and reliability under tight deadlines.
• Mission alignment with SecureBio’s goal of preventing catastrophic pandemics.

Responsibilities

• Run SecureBio’s panel of biosecurity evaluations on frontier AI models using Inspect AI, an open-source Python framework.
• Identify opportunities to improve automation, efficiency, and robustness of the evaluation pipeline.
• Develop and improve evaluation code, working with other software engineers to maintain quality, reproducibility, and scalability.
• Support interpretation and ensuring accuracy of eval results by collaborating with evaluation developers and researchers.
• Draft report outlines to facilitate analyses and communication to AI firms and other audiences.
• Develop data visualizations and figures for technical reports and dashboards.

Skills

Python

Software engineering

Statistics

Data visualization

Large language models

Codebase navigation

Communication

Education

Background in life sciences

Tools

Docker

Web APIs

Inspect AI evaluation framework

Frontend development

Job description

• Project/Division: AIxBio
• Start date: As soon as possible
• Status: Full-time, (contract-to-hire considered)
• Approximate compensation: $70,000-$115,000 per year, based on experience. This range assumes a location in the United States; there would be adjustments downwards for candidates based in most other geographic locations. Higher compensation may be considered for exceptional candidates.
• Work location: Remote
• Reports to: Research Scientists

The role

SecureBio seeks a Software Engineer to help scale and systematize the way we run evaluations of frontier AI systems, with a focus on biosecurity and misuse risk. This work has already made an impact: our biosecurity evaluations have been publicly cited in model cards from OpenAI, Anthropic, and Google, and featured in xAI's Risk Management Framework.

As a Software Engineer, you will own running our evaluation suite on new and emerging AI models. This includes ensuring that mature evaluations are correctly configured and run through our Inspect platform, maintaining and updating our dashboard of results, and providing structured results for written analyses and reports.

This role is an opportunity to contribute directly to SecureBio’s mission of reducing catastrophic biological risks by ensuring that our biosecurity assessments of frontier models are robust, reproducible, and timely. Over time, you will gain experience across our evaluation ecosystem and play a key role in maintaining reliable systems that scale with new models and collaborations.

Responsibilities

• Running evaluations
  • Run SecureBio’s panel of biosecurity evaluations on frontier AI models using Inspect AI, an open-source Python framework, for both pre-release testing and SecureBio’s web dashboard.
  • Identify opportunities to improve automation, efficiency, and robustness of the evaluation pipeline.
• Refining evaluations
  • Flag anomalies, unexpected model behaviors, and potential scoring issues for expert review.
  • Develop and improve evaluation code, working collaboratively with other software engineers to maintain code quality, reproducibility, scalability, and documentation.
  • Support interpretation and ensuring accuracy of eval results by working closely with evaluation developers, research scientists, and subject-matter experts.
• Supporting communication
  • Draft report outlines to facilitate analyses and communication to AI firms and other audiences. In many cases, SecureBio’s safety testing directly informs AI companies’ safety policies and publications (such as model system cards).
  • Develop data visualizations and figures for technical reports, stakeholder communications, and for SecureBio’s web dashboard of evaluation results.

Representative Projects

• Perform pre-release safety testing on frontier AI models using SecureBio’s biosecurity evaluation suite.
• Contribute to SecureBio’s research reports and stakeholder communications on AI-enabled biorisk by producing structured outputs and draft analyses.
• Maintain, populate, and improve SecureBio’s web dashboard and associated results database of biosecurity-relevant AI evaluations.Collaborate with other software developers to improve existing evaluations and develop new evaluations.

Qualifications

Required

• 1-2 years of professional software engineering experience, particularly with Python.
• Familiarity with basic statistics, data analysis, and data visualization.
• Proficiency in using large language models in personal or professional contexts.
• Demonstrated ability to navigate and understand unfamiliar codebases.
• Strong communication skills, responsiveness, attention to detail, and reliability under tight deadlines.
• Mission alignment with SecureBio’s goal of preventing catastrophic pandemics.

Preferred

• Experience with Docker and web APIs.
• Experience with the Inspect AI evaluation framework.
• Experience with frontend web development.
• Background in life sciences (e.g. undergraduate degree in biology, or biotech work experience).
• Demonstrated ability to analyze and troubleshoot the behavior of complex software systems.
• Interest in building long-term expertise in evaluation and model assessment.

Salary and benefits

This is a full-time, remote role, reporting to Research Scientists. The starting salary range is $70,000 to $115,000 per annum, depending on prior experience. This range assumes a location in the United States; there would be adjustments downwards for candidates based in most other geographic locations. There may be flexibility in salary for exceptional candidates with significant relevant experience.

Our employee benefits reflect our belief in investing to build the strongest possible team, in which everyone receives the trust and support they need to excel. Our benefits package includes:

About SecureBio

In joining SecureBio, you'll be joining a motivated, mission-driven, and international team of professionals dedicated to protecting the world from biological threats. Through in-house research & development, as well as collaborations with experts in academia, industry & government, we are working to build the tools, technologies, and institutions the world needs to be secure against future biotechnology,

Major projects by the SecureBio team include the Nucleic Acid Observatory project for reliable pandemic detection and our work to understand and reduce biological risks from advanced artificial intelligence.

Referral bonus

We offer a referral bonus of up to $3,000 to individuals who refer successful candidates for this position:

• If you refer a candidate that we invite to a full structured interview or work trial, you will receive $350;
• If the candidate accepts a job offer, you will receive a further $1,150, for a total of $1,500;
• If the candidate is still working with us after the three months from their start date, you will receive a further $1,500, for a total of $3,000.

To receive these bonuses, we must reasonably believe that you were responsible for referring the candidate to us (e.g. due to you introducing them to us when we didn’t previously know them, or them mentioning you on their application as the reason they heard about the position).