FirstPrinciples Research Fellow

FirstPrinciples is a non-profit organization building an autonomous AI Physicist to understand the nature of reality: the underlying structure, governing principles, and fundamental laws of our universe. We're developing an intelligent system that can explore theoretical frameworks, reason across disciplines, and generate novel insights to tackle the deepest unsolved problems in physics. By combining AI, symbolic reasoning, and autonomous research capabilities, we're developing a platform that goes beyond analyzing existing knowledge to actively contribute to physics research. Our goal is to accelerate progress on the questions that have captivated humanity for centuries.

We operate as a global nonprofit organization, with a Canadian foundation, a US-based 501(c)(3).

As part of this effort, we are launching a Research Fellowship Program for advanced AI and physics researchers. Fellows will work directly with the FirstPrinciples Research and Engineering teams to design, test, and implement state-of-the-art (SOTA) methods and applications that will be integrated into the core AI Physicist system. This is not a paper-only fellowship and your work will go straight into production, helping shape how scientific research is performed.

What You will Do As a Research Fellow, you will own a well-scoped research direction that contributes directly to the AI Physicist's ability to reason about physics.

Core Objectives:

Improve Theo’s ability to produce scientifically sound, high-quality outputs in 2026;

Introduce new ideas, methods, and approaches that meaningfully shift system performance;

Bring deep domain expertise (PhD+ level) in one or more targeted research areas.

Research Areas & Responsibilities

Fundamental Model Research:

Research, design, and test novel model architectures that combine academic literature, NLP, symbolic reasoning, and structured scientific workflows.

Prototype and build embedding representations for physical concepts, mathematical objects, and logical structures, enabling models to reason over equations, abstractions, and scientific constraints rather than surface text alone.

Investigate alternatives to transformer-based architectures and deliver concrete recommendations.

Design and run targeted experiments to evaluate new architectural ideas, using empirical results to guide the development of next-generation model architectures.

Develop reinforcement learning loops that enable models to run internal and independent thought experiment.

Multimodal Data & Benchmarking:

Design and automate scalable data ingestion pipelines that aggregate scientific literature, metadata, equations, and experimental data.

Create custom benchmarks to measure physical understanding, mathematical reasoning, and failure modes in scientific reasoning and abstraction.

Refine and release curated datasets and baselines once internal validation is complete.

Training, Testing & Safety:

Run and track model training jobs while managing compute usage and budget constraints.

Design sandbox environments for controlled autonomous exploration.

Build evaluation frameworks using visual and statistical tools to identify strengths and blind spots.

Implement tests and guardrails that flag low-quality or unsafe outputs.

Maintain internal issue tracking with clear failure modes and fixes.

Collaboration & Technical Guidance:

Work closely with engineers to ensure research is feasible and production-ready.

Communicate technical trade-offs clearly to non-technical stakeholders.

Present regular research updates tied to defined milestones.

Who This Fellowship Is For We are looking for researchers who:

Can operate independently and own a problem end-to-end.

Are motivated by ambitious, long-horizon goals.

Have strong builder instincts, not just theory.

Value rigor, clarity, and intellectual honesty.

Are comfortable engaging with exploratory, incomplete results.

Typical profiles include:

PhD or postdoctoral researchers in Computer Science, Machine Learning, Theoretical Physics, or a closely related field.

Track record of research in either:

model architectures, representation learning, or reasoning systems (CS/ML path); or

mathematical, physical, or formal reasoning applied to fundamental problems (physics path).

Demonstrated ability to translate abstract theory into testable computational systems.

Comfortable working across disciplinary boundaries.

Program Structure

6-12 month engagement with full-time or equivalent commitment

Fully remote

Fellows participate in projects and are expected to produce outputs (code, models, and other deliverables)

Fixed stipend for the full term

How to Express Interest: If this resonates, we welcome an expression of interest, including:

Resume;

Description of your research background;

Primary areas of expertise; and

Relevant recent work.

Join us at FirstPrinciples and be a part of a transformative journey where science drives progress and unlocks the potential of humanity.