Applied Scientist

Role overview Develop and validate scientific methods that reveal what drives outcomes in complex, unstructured evidence. The work combines statistical and causal inference with machine learning and large language models, moving from experiments and prototypes to Python pipelines used in practical analysis. You’ll collaborate with engineers and decision-makers to make results rigorous, scalable, and useful.

Responsibilities - Design and test methods for extracting reliable insights from internal text and other unstructured evidence. - Explore how statistical inference, causal frameworks, machine learning, and language models can work together. - Prototype analytical approaches and implement promising ones as well-designed Python pipelines. - Evaluate and benchmark statistical, machine learning, and AI outputs; investigate where models fail. - Work with engineers and client teams to ensure analyses meet real-world needs and hold up to scrutiny. - Help shape the scientific methodology and practices of an early-stage team.

Requirements - Strong grounding in statistical inference, causal inference, and machine learning. - Experience turning analytical ideas into Python implementations. - Practical experience experimenting with large language models beyond basic use. - Substantial experience evaluating or benchmarking statistical, machine learning, or AI systems. - Interest in understanding cause and effect in messy, real-world data.

Nice to have - Analysis experience that informed consequential decisions under scrutiny or time pressure. - Interest in human judgment, behavior, or decision-making. - Familiarity with data collection and sampling methods. - Experience working in an early-stage or startup environment.