Member of Technical Staff, Quantitative Research

Testudo•New York, NY

5d•$150,000 - $250,000•Onsite

About The Position

As a Member of Technical Staff, Quantitative Research at Testudo, you'll develop quantitative models for some of the hardest risks created by artificial intelligence, autonomous systems, and other emerging technologies. You'll be given difficult, ambiguous quantitative questions where the data may not exist and there may be no established methodology. You will reason about probability and severity from first principles, form a thesis, determine what evidence is required, source or generate the data, and build and validate the models needed to reach a robust answer.

Requirements

Exceptional quantitative ability, with deep expertise in statistics, probability, machine learning, applied mathematics, econometrics, or a similarly rigorous discipline.
A track record of independently solving difficult, ambiguous quantitative problems — for example as a quantitative researcher at a hedge fund or market maker, an ML researcher, a statistician, or through advanced academic research such as a PhD in a related field.
Can reason from first principles about the probability and severity of poorly understood events, and translate that reasoning into quantitative models.
Can source or generate your own data when the dataset you need does not already exist.
Deeply comfortable with probability distributions, uncertainty, sparse evidence, noisy signals, and estimates where confidence may necessarily be limited.
Can independently write production-quality code, contribute directly to Testudo's codebase, and take quantitative models from research through implementation.
Think creatively about numerical problems and are willing to invent new approaches when established methods are insufficient.

Nice To Haves

Need a clean, labelled dataset before you can start.
Prefer publication cycles to shipping models into production.
Want an established methodology to apply rather than one to invent.
Prefer remote or hybrid.

Responsibilities

Estimating the probability and severity of frontier risks, such as an autonomous vehicle causing an accident or an AI system providing materially incorrect professional advice.
Building frequency and severity models from scratch for risks where no established approach exists, and handing them to our actuarial team to price against.
Developing statistical and machine learning models where historical data is sparse and noisy.
Designing experiments and large-scale data generation and LLM labeling pipelines to create the datasets required for modeling.
Building probability distributions, simulations, and uncertainty estimates for rapidly evolving technologies.
Taking quantitative problems end-to-end: from thesis formation and data collection through modeling, validation, and implementation.
Turning quantitative research into production pricing systems and contributing directly to Testudo's codebase.