Research Scientist

present of AI; Australia; Remote; AUD 120,000 - 180,000; Full time

Job summary

Metacognition AI seeks a research scientist to develop and validate questions, run projects to completion, and publish at top AI conferences. A PhD with a strong publication record is expected, with deep knowledge in reasoning, memory, multimodal learning, or training large models.

The role includes guiding PhD interns and junior researchers, with scope tailored to the background. Initially you will shape a small research program and mentor a small team, while influencing direction of core AI

Qualifications

• PhD with a strong publication record in AI.
• Expertise in reasoning, memory, multimodal learning, or large-model training.
• Experience guiding PhD interns and junior researchers.

Responsibilities

• Develop research questions and drive projects to completion.
• Publish results at top AI conferences.
• Mentor interns and junior researchers.

Skills

Reasoning

Memory

Multimodal learning

Research management

Education

PhD in AI/ML

Job description

This is a research job for someone who can pick good questions, run the work to completion, and get it accepted at top AI conferences. A strong candidate holds a PhD, has a solid record of published papers, and has deep knowledge in areas like reasoning, memory, multimodal learning, or training large models. Helping guide PhD interns and junior researchers is part of the work, and the scope will be set to fit the person's background.

How it fits the AI era

The work sits at the core of the AI field, so it is about advancing the models and systems themselves rather than using AI as a tool. Day to day, AI helps with coding experiments, reading papers and testing ideas faster. What stays human is choosing which questions matter, judging whether results are sound, and guiding others through the research.

How it builds AI experience

The role offers hands‑on research into how AI agents remember, reason and plan, which fits a company building a base layer that controls what agents can do. Someone in this seat could gain experience linking research ideas to an open source agent platform. To build toward it, follow open agent software, try memory and planning experiments on small models, and write up findings clearly.

Worth showing

• Independent research from idea to paper
• Reasoning or memory expertise

This is our own read, written from the employer's posting and the company's Spotlight. The employer's posting is the source of truth for the role.

About Metacognition AI

An operating system for AI agents that sits beneath language models to control what they can do, released as open source.