Research Scientist/Engineer (Evaluations)

Location: In-person role based in the London or San Francisco office; flexible work-from-home arrangements are offered.

Salary: £100,000–£200,000 per year (approximately $150,000–$270,000 USD).

Company

hirify.global studies risks from loss of control in advanced AI systems, with a focus on deceptive alignment and scheming.

What you will do

• Run and own pre-deployment evaluation engagements with frontier AI labs, conducting thousands of runs across hundreds of environments.
• Design and build automated pipelines for training-run assessments and post-training data analysis.
• Evaluate and red-team model checkpoints at different stages of post-training.
• Investigate alignment faking, scheming, metagaming, and other undesirable model behaviors through targeted experiments.
• Develop new evaluations and infrastructure for automated red-teaming between evaluation campaigns.
• Communicate quantitative and qualitative findings to technical and non-technical audiences and help translate results into recommendations.

Requirements

• Strong software engineering skills and experience with production Python code.
• Ability to extract meaningful signals from large, messy datasets through quantitative analysis and qualitative assessment.
• Ability to identify anomalies and unexpected model behaviors.
• Clear, concise technical and non-technical writing and communication skills.
• Experience using AI tools effectively and evaluating when different models are appropriate.
• Willingness to work full-time from the London or San Francisco office, with flexible work-from-home arrangements available.

Nice to have

• Experience thinking about AI risk topics such as scheming and metagaming.
• Knowledge of LLM post-training, including RLHF, reasoning training, supervised fine-tuning, and on-policy distillation.
• Experience creating evaluations with Inspect, Harbor, or similar frameworks.
• Self-taught or non-traditional background; formal industry experience is not required.

Culture & Benefits

• Market-competitive salary, equity, and benefits.
• Flexible working hours and schedule, unlimited vacation, and unlimited sick leave.
• Up to six months of paid parental leave and comprehensive health, dental, and vision insurance.
• Retirement savings with competitive employer matching, including 401(k) for US employees.
• Paid work trips, staff retreats, relevant conferences, and a $1,000 annual professional development budget.
• Relocation support and visa fees may be provided; visa sponsorship is available for the UK and US, subject to the role and candidate.

Hiring process

• Submit a CV; a cover letter is optional, and relevant work samples may be included.
• Complete a screening interview, an approximately 2.5-hour take-home test, three technical interviews, and a final interview with the CEO.
• Technical interviews focus on practical tasks related to the role rather than general LeetCode-style coding questions.