Remote AI Researcher (PhD/Master) - Advanced Reasoning

Alignerr Corp.; Oxford; Remote; GBP 41,000 - 93,000; Part time

Job summary

Alignerr is seeking Masters and PhD‑level researchers to advance AI training by designing and evaluating complex problems that push reasoning to a graduate level. This remote, hourly contract role offers autonomy and intellectual challenge for researchers who value rigorous science and flexible schedules.

You will develop multi-step problem statements, review AI-generated solutions for rigor, and collaborate asynchronously with other domain experts to improve model performance while maintaining

Qualifications

• Masters or PhD in a STEM or highly technical field from a leading university.
• Strong track record in research, peer‑reviewed publication, or complex technical project management.
• Exceptional written communication skills and rigorous attention to detail.

Responsibilities

• Design and develop complex, multi-step problems across your domain that challenge and sharpen AI reasoning.
• Apply your research expertise to craft problem statements with real academic depth and real-world relevance.
• Review and evaluate AI-generated solutions for correctness, rigor, and conceptual accuracy.
• Collaborate asynchronously with AI researchers and domain experts to improve model performance.
• Ensure every deliverable meets a high standard of clarity, precision, and academic integrity.

Skills

Research experience

Academic writing

Self-directed work

Asynchronous collaboration

Attention to detail

Python

R

Education

Master's or PhD in STEM or technical field

Tools

Python

R

Job description

Alignerr is seeking Masters and PhD‑level researchers to advance AI training by designing and evaluating complex problems that push reasoning to a graduate level. This remote, hourly contract role offers autonomy and intellectual challenge for researchers who value rigorous science and flexible schedules.

You will develop multi-step problem statements, review AI-generated solutions for rigor, and collaborate asynchronously with other domain experts to improve model performance while maintaining