Staff Research Scientist

Turing is recruiting a Staff Research Scientist to study how frontier AI systems are judged, trained and made dependable. The listing marks the position on-site, and no city is named. The work tackles questions where suitable benchmarks, datasets or methods do not yet exist, so the scientist spots gaps in the literature, proposes ambitious programmes, and carries projects from first idea through experiments, benchmark construction and publication.

Four research strands are named. Frontier evaluation covers building benchmarks across STEM disciplines and general model skills, along with methods for generating tasks, grading answers, preventing contamination and calibrating difficulty. Synthetic data and post-training examines how to produce strong STEM training material, and when it truly lifts capability rather than merely adding volume. Reliability work studies hallucination, uncertainty, calibration and factual accuracy, including when a model should reason internally, use tools or decline to answer. Agentic science looks at long-horizon agents that undertake extended technical work, and at what keeps them from doing that reliably. Proposals in reasoning, AI for science and new capabilities are also welcome.

The posting calls for a doctorate or comparable research record in biology, chemistry, statistics, engineering, mathematics, physics, computing, machine learning or any other technical discipline. The scientist should be able to frame and run original studies, understand current large language models in depth, design sound experiments, and absorb unfamiliar literature quickly. Python is needed for prototypes and evaluation pipelines, and writing must be precise. A publication record is valued, but the employer weighs most heavily the ability to pick important questions and answer them rigorously and promptly.

The company says it works with frontier labs, academic researchers and enterprises, and publishes its findings, datasets and benchmarks openly. Listed compensation is $250,000 to $400,000 on-target earnings plus equity.