Founding Research Scientist, Human Simulation

Listen Labs•San Francisco, CA

About The Position

Listen is building the human layer of AI, a preference model trained on millions of real human conversations. This role will lead the simulation initiative, which aims to create a model that allows AI systems to predict what humans would think, want, and decide. The company is Sequoia-backed, has raised $100M, and serves customers like Anthropic and Google. Listen's platform runs AI-moderated video interviews at scale, enabling rapid insights into what customers want. The future vision is to expose this human simulation capability as a 'Human API' for AI agents to use in product development, enabling closed-loop iteration at AI speed.

Requirements

Strong research track record with published work in LLMs, post-training, RLHF, behavioral modeling, simulation, or adjacent fields, or equivalent industrial impact at a frontier lab.
Ability to define research problems and scope research programs.
Genuine curiosity about human behavior, decision-making, and preference shifts.
Ability to train models, write evaluations, and collaborate with research engineers.
Excellent written communication skills for conveying complex ideas and roadmaps.

Nice To Haves

Experience in multi-agent dynamics and simulating group behavior.
Experience with generalization and active learning in large-scale data.
Experience in understanding how to combine long-form interviews, demographics, and behavior into a useful model.
Ability to estimate confidence in predictions.

Responsibilities

Lead the human simulation initiative.
Define research programs and determine success metrics.
Train models, write evaluations, and collaborate with research engineers to deploy models into production.
Communicate complex ideas and the roadmap for the initiative in writing.