Apollo Research Opens Research Scientist/Engineer (Evaluations) Position in London and San Francisco @Unknown

Apollo Research is inviting applications for the Research Scientist/Engineer (Evaluations) position, offering an exciting opportunity for professionals passionate about frontier artificial intelligence, AI safety, machine learning evaluation, and large language models (LLMs). The full-time, on-site role is available in London, United Kingdom, and San Francisco, California, with visa sponsorship available for eligible candidates.

The organisation is actively recruiting and conducting interviews on a rolling basis, with the aim of filling the position as soon as a suitable candidate is identified.

Successful applicants will join Apollo Research’s Pre-Deployment Team, where they will help design and implement advanced evaluation systems to assess the behaviour and safety of cutting-edge AI models before public deployment.

About Apollo Research

Apollo Research is an independent AI safety research organisation dedicated to reducing risks associated with increasingly capable artificial intelligence systems. Its work primarily focuses on mitigating Loss of Control risks, particularly scenarios where advanced AI models may behave deceptively or develop misaligned objectives that evade human oversight.

The organisation collaborates closely with leading frontier AI laboratories, including OpenAI, Anthropic, and Google DeepMind, helping evaluate new AI models before they are released.

Apollo Research conducts research across several core areas, including:

• AI model evaluations.
• Automated red-teaming.
• Detection of deceptive alignment.
• Research into AI scheming behaviours.
• AI control and mitigation strategies.
• Safety evaluation methodologies.
• Frontier AI research collaboration.

About the Research Scientist/Engineer Role

The successful candidate will contribute to Training-Run Assessments (TRAs), an important component of Apollo Research’s pre-deployment safety evaluations.

The role focuses on building automated systems capable of detecting undesirable behaviours emerging during the post-training phase of frontier AI models.

Researchers will work with newly developed models before public release, conducting rigorous evaluations to determine whether behaviours such as deceptive alignment, alignment faking, or other forms of model misalignment are present.

Key Responsibilities

The Research Scientist/Engineer will contribute to both operational evaluation campaigns and ongoing research methodology development.

Primary responsibilities include:

• Leading and managing pre-deployment AI evaluation campaigns.
• Designing automated assessment pipelines for frontier AI systems.
• Evaluating model checkpoints throughout post-training stages.
• Conducting large-scale AI red-teaming exercises.
• Investigating behaviours learned during AI training.
• Detecting undesirable model behaviours, including alignment faking and scheming.
• Analysing post-training datasets.
• Producing technical reports for partner AI laboratories.
• Improving automated evaluation methodologies between assessment campaigns.
• Building infrastructure that supports scalable AI evaluations.

The team typically conducts comprehensive evaluation campaigns approximately every two weeks, involving thousands of testing runs across hundreds of unique environments.

Required Qualifications and Skills

Apollo Research welcomes applicants from both traditional and non-traditional backgrounds.

The organisation does not require formal industry experience and encourages applications from self-taught professionals who possess strong technical capabilities.

Applicants should demonstrate:

• Strong software engineering skills using Python.
• Experience developing and maintaining production-quality Python code.
• Ability to design clean, reusable software architectures.
• Experience analysing large and complex datasets.
• Strong quantitative and qualitative analytical skills.
• Ability to identify unexpected AI behaviours and anomalies.
• Excellent written and verbal communication skills.
• Ability to explain technical concepts to both technical and non-technical audiences.
• Practical experience using multiple AI models to improve productivity and research workflows.
• Curiosity and willingness to experiment with emerging AI technologies.

Preferred Qualifications

While not mandatory, candidates with experience in the following areas are particularly encouraged to apply:

• AI safety research.
• Large Language Model evaluation.
• Reinforcement Learning from Human Feedback (RLHF).
• Supervised fine-tuning.
• On-policy distillation.
• AI reasoning training.
• AI red-teaming methodologies.
• Inspect evaluation framework.
• Harbour or similar AI evaluation platforms.

Apollo Research emphasises that applicants who may not meet every listed qualification but believe they are a strong fit should still submit an application.

Salary and Employee Benefits

Apollo Research offers a highly competitive compensation package designed to attract exceptional AI researchers and engineers.

Benefits include:

• Annual salary ranging from £100,000 to £200,000.
• Equity participation.
• Flexible working hours.
• Unlimited annual leave.
• Unlimited sick leave.
• Up to six months of paid parental leave.
• Comprehensive medical, dental, and vision insurance.
• Retirement savings plans with employer contributions.
• Complimentary breakfast, lunch, dinner, and snacks during office workdays.
• Fully funded staff retreats, conferences, and business travel.
• Annual professional development budget of USD 1,000.
• Relocation assistance.
• Visa sponsorship for eligible candidates in both the United Kingdom and United States.

Working Environment

The position is primarily based in Apollo Research’s offices in London or San Francisco.

Although the organisation supports flexible working hours and work-from-home arrangements where appropriate, this is fundamentally an in-person role designed to encourage close collaboration among researchers, engineers, and AI safety specialists.

Successful candidates will work alongside internationally recognised researchers specialising in AI evaluation, governance, alignment research, and machine learning safety.

Recruitment Process

Apollo Research follows a structured multi-stage recruitment process designed to evaluate practical research and engineering skills.

The selection process includes:

• Initial screening interview.
• Approximately 2.5-hour take-home assessment.
• Three technical interviews.
• Final interview with the Chief Executive Officer.

Unlike many software engineering recruitment processes, Apollo Research does not use traditional algorithm-focused coding interviews. Instead, technical interviews focus on practical tasks directly related to AI evaluation and research.

Applicants interested in strengthening their applications are encouraged to build hands-on Large Language Model evaluation projects using frameworks such as Inspect before interviewing.

Commitment to Diversity, Fairness, and Responsible AI

Apollo Research is an Equal Opportunity Employer committed to creating an inclusive workplace where individuals from diverse backgrounds can thrive.

Employment decisions are made without discrimination based on race, gender, disability, religion, age, sexual orientation, pregnancy, marital status, or other protected characteristics.

The organisation also promotes responsible artificial intelligence within its recruitment process. Although AI-assisted tools may support administrative tasks such as resume screening, all applications are reviewed by human recruiters, and final hiring decisions are made by Apollo Research’s recruitment team.

Why Apply?

The Research Scientist/Engineer (Evaluations) role offers a unique opportunity to contribute directly to the safety of frontier artificial intelligence systems before they are deployed globally. Researchers will collaborate with leading AI laboratories, work on some of the world’s most advanced language models, and help shape the future of responsible AI development through rigorous evaluation, red-teaming, and safety research.

Professionals with expertise in software engineering, machine learning, data analysis, AI safety, and LLM evaluation who are passionate about advancing trustworthy artificial intelligence are encouraged to apply while recruitment remains open.

VISIT OFFICIAL WEBSITE TO APPLY

For more opportunities such as these please follow us on Facebook, Instagram, Twitter, LinkedIn and WPChannel

Disclaimer: Global South Opportunities (GSO) is not the organization offering this opportunity. For any inquiries, please contact the official organization directly. Please do not send your applications & CVs to GSO, as we are unable to process them. Due to the high volume of emails, we receive daily, we may not be able to respond to all inquiries. Thank you for your understanding.