Bioinformatics Engineer - RefusalBench

 Posted 2 hours ago
     
 $120K - $180K per year
  
5-10 years experience
Apply Now

Please mention DailyRemote when applying

AI Summary

Review biological tasks and research workflows to determine biosecurity risks and establish ground truth for AI evaluations. Convert real-world scientific examples into structured evaluations to test AI model refusal capabilities.

About the Role

At LatchBio, we build the benchmarks that frontier AI labs use to evaluate and train models on biological reasoning. RefusalBench tests whether AI systems can distinguish legitimate biological research from requests that present meaningful biosecurity risks.

We're looking for scientists with deep expertise in areas such as biosecurity, biosafety, pathogen genomics, infectious disease research, synthetic biology, biodefense, and public health. Your role will be to apply that expertise to determine where the boundary lies between routine scientific work and potentially dangerous biological capabilities.

You will review real-world biological analyses, protocols, datasets, and research papers to establish ground truth for AI evaluations. Some tasks should clearly be allowed. Others should clearly be refused. Many sit in the gray area. Your job is to identify the relevant risks, justify the correct decision, and convert those examples into structured evaluations that test whether AI systems reach the same conclusion.

What You'll Do

  • Review biological tasks, datasets, protocols, and research workflows to determine whether they should be accepted or refused from a biosecurity perspective

  • Identify the specific risk categories involved, including pathogen engineering, immune evasion, enhanced transmissibility, countermeasure resistance, and other dual-use concern

  • Source real-world examples from the scientific literature and convert them into evaluation tasks with clear ground truth and supporting rational

  • Document the reasoning and evidence supporting each decision so that evaluations are scientifically defensible and consistently graded

  • Review agent outputs to determine whether models correctly identify risks, over-refuse legitimate research, or fail to recognize dangerous capabilities

  • Help build a high-quality corpus of biosecurity evaluations grounded in real biological research and real-world risk assessment

Ideal Backgrounds

  • Biosecurity & biodefense programs

  • Biosafety and risk assessment organizations

  • Public health agencies (CDC, WHO, state health departments)

  • Pathogen genomics and molecular epidemiology groups

  • Synthetic biology and biotechnology governance

  • Pandemic preparedness and emerging infectious disease programs

  • Infectious disease surveillance and outbreak response

Compensation & Logistics

  • Salary: $120k–$180k (base + performance pay that scales with output)

  • Produce 50 evals/week at baseline — exceed that and performance pay scales linearly

  • Fast-track promotion for strong performers

  • Fully remote or onsite

  • Equity

  • 100% premium-covered Blue Shield Platinum health plan ($0/$0)

  • Unlimited PTO

  • Visa sponsorship available

Hiring Process

  • Intro call with Saúl (Technical Recruiter)

  • Technical interview with Harmon (Pod Lead)

  • Cultural interview with Jordan (Chief of Staff)

  • Offer

Location: San Francisco, CA. In-person.

Similar Jobs

See all Remote Software Development jobs →

Personalize your Remote Job Search in 3 Easy Steps!

Discover remote opportunities in Software Development

Answer easy questions

Answer easy questions

200,000+ jobs across 15+ categories

Get your best job matches

Get your best job matches

Only hand-screened, legit jobs

Find a remote job faster

Find a remote job faster

No ads, scams, or junk

I was the first applicant for a remote marketing position that got listed on the company website the same day I applied. Had an interview within 48 hours!

Sarah J. — Sarah J. · Marketing Manager ★★★★★ Verified