Application Guide
How to Apply for Research Scientist/Engineer (Evaluations)
at Apollo Research
🏢 About Apollo Research
Apollo Research is uniquely focused on one of the most critical AI safety problems: detecting and mitigating deceptive alignment and scheming in frontier AI systems. Unlike general AI safety organizations, they specialize specifically in evaluations and collaborate directly with leading AI labs to test the most advanced models before deployment. Working here means tackling concrete, high-stakes problems at the cutting edge of AI risk research with direct access to state-of-the-art systems.
About This Role
This Research Scientist/Engineer (Evaluations) role involves designing and running pre-deployment evaluation campaigns on the world's most capable AI models to detect deceptive behaviors and frontier risks. You'll analyze thousands of model interactions to identify patterns, build novel test environments, and automate evaluation pipelines. This work directly impacts AI safety by providing critical detection capabilities before potentially dangerous systems are deployed.
💡 A Day in the Life
A typical day might involve running evaluation campaigns on the latest frontier models from partner labs, analyzing transcripts to identify behavioral patterns, and improving automated testing pipelines. You could spend the morning designing novel test scenarios for deceptive alignment detection, the afternoon scaling evaluations across hundreds of scenarios, and the evening documenting findings for technical reports and stakeholder presentations.
🚀 Application Tools
🎯 Who Apollo Research Is Looking For
- A Python engineer who has shipped production code and can optimize complex evaluation pipelines, not just write scripts
- Someone with exceptional pattern recognition skills who enjoys digging through messy datasets of model transcripts to surface behavioral insights
- An AI power-user who understands both the technical details of model behavior and can communicate findings clearly to diverse audiences
- A process optimizer who actively looks for friction points in workflows and enjoys building scalable, automated solutions
📝 Tips for Applying to Apollo Research
Showcase specific examples of Python projects you've shipped to production, emphasizing code quality and maintenance experience
Demonstrate your ability to extract insights from messy data by describing a complex dataset you analyzed and what patterns you discovered
Highlight any experience with AI model evaluation, testing frameworks, or working with large language model APIs
Explain how you've optimized workflows in past roles, quantifying the efficiency gains you achieved
Show familiarity with Apollo's specific research focus by referencing their work on deceptive alignment or scheming detection
✉️ What to Emphasize in Your Cover Letter
["Your experience with production Python code and specific examples of software you've shipped and maintained", "How you've extracted meaningful patterns from large, unstructured datasets in past projects", 'Your understanding of AI safety risks, particularly deceptive alignment/scheming, and why evaluation is critical', 'Examples of workflow optimization where you reduced friction and improved efficiency in technical processes']
Generate Cover Letter →🔍 Research Before Applying
To stand out, make sure you've researched:
- → Read Apollo's published research on deceptive alignment and scheming to understand their specific technical approach
- → Review their collaborations with frontier AI companies to understand their access model and partnerships
- → Study their evaluation methodology papers to understand their current technical approaches
- → Look into their team members' backgrounds to understand the interdisciplinary nature of their work
💬 Prepare for These Interview Topics
Based on this role, you may be asked about:
⚠️ Common Mistakes to Avoid
- Focusing only on theoretical AI safety without demonstrating practical engineering and evaluation skills
- Presenting generic data analysis experience without showing ability to work with messy, unstructured text data from AI interactions
- Applying with a generic AI/ML background without tailoring your application to Apollo's specific focus on evaluations and frontier risk detection
📅 Application Timeline
This position is open until filled. However, we recommend applying as soon as possible as roles at mission-driven organizations tend to fill quickly.
Typical hiring timeline:
Application Review
1-2 weeks
Initial Screening
Phone call or written assessment
Interviews
1-2 rounds, usually virtual
Offer
Congratulations!
Ready to Apply?
Good luck with your application to Apollo Research!