PhD Researcher — AI & Human Behavior… at Squoosh.AI · Remote

PhD Researcher — AI & Human Behavior (Part-Time)

Squoosh.AI | Remote within the United States | $60–70/hour USD

Continue your doctorate while owning research that becomes a real product. Squoosh.AI is hiring a current full-time PhD student to work part-time with our research team. We will agree on a manageable weekly schedule around your academic commitments.

Squoosh develops AI systems that simulate how people explore and interact with digital experiences, helping commerce teams evaluate and improve websites. Our research connects machine learning, agent behavior, human decision-making, and rigorous experimentation.

What you will work on

Build and evaluate LLM-based browser agents that navigate websites, use tools, and make decisions under realistic constraints.
Investigate persona calibration and behavioral modeling: how goals, preferences, context, and individual differences affect simulated journeys.
Design benchmarks, baselines, ablations, and error analyses that compare simulations with observed behavior and real A/B-test outcomes.
Explore experimental design, causal inference, uncertainty, and generalization to understand when simulations support useful product decisions.
Write reproducible Python experiments and collaborate with researchers and engineers to turn useful findings into prototypes.

What we are looking for

Current full-time enrollment in a relevant PhD program: AI, computer science, machine learning, statistics, HCI, cognitive science, computational social science, or a related field.
Strong Python implementation skills and a solid foundation in statistics and research methods.
A research artifact you can explain in depth: a paper, preprint, repository, benchmark, or experimental write-up. Publication is helpful, not required.
Clear thinking about baselines, data leakage, failure cases, and what your evidence does and does not support.
The ability to own a focused question and communicate progress while maintaining a reliable part-time commitment.

Useful experience includes PyTorch or JAX, LLMs and agents, reinforcement learning, browser automation, synthetic data, model evaluation, and human behavior research. Deep ability in one relevant area matters more than familiarity with every tool.

How to apply

Apply with your academic CV. Include your program and research area, availability and expected graduation date, one or two shareable research artifacts, your personal contribution, and a short example of an experiment and what its failure cases taught you.

Your doctoral studies remain your primary academic commitment. This is separate part-time employment, not admission to a PhD program. US work authorization is required; applicable CPT/OPT candidates are welcome. Only share material you are permitted to share. Squoosh.AI is an equal opportunity employer.