About this role
This role addresses the need to advance the capabilities of frontier AI systems by focusing on rigorous evaluation, data generation, and the development of agentic science within STEM domains. Turing operates at the intersection of frontier AI research and real-world enterprise deployment, aiming to close the loop between research and deployment by generating better data and more capable models. The Research Scientist will be instrumental in identifying critical gaps in current literature concerning STEM evaluation and developing novel methodologies to address them.
The individual will be responsible for transforming initial research hypotheses into concrete, executable projects, spanning from hypothesis formulation through experimentation, benchmark construction, and eventual publication. Key areas of focus include creating new benchmarks, developing synthetic data generation techniques, and investigating reliability and hallucination mechanisms in technical reasoning.
This position is ideal for a scientist who thrives on defining problems where established solutions are absent and possesses the drive to design and execute complex, multi-stage research programs. Success in this role requires a capacity to synthesize complex technical literature quickly and apply strong scientific judgment to ambiguous, high-impact questions.
What you will do
- Identify significant shortcomings in existing benchmark and evaluation literature across STEM fields and general model functionality.
- Design and construct novel benchmarks tailored to evaluate emerging model capabilities that static benchmarks fail to capture.
- Develop rigorous methodologies for task generation, grading, contamination control, and difficulty calibration.
- Create methods for generating high-quality synthetic STEM training data and assess its impact on downstream performance.
- Investigate issues related to hallucination, uncertainty, and factual reliability in technical domains.
- Research and develop workflows for agentic science, focusing on extended scientific work involving literature search and tool use.
- Evaluate long-horizon scientific agents to identify bottlenecks in their ability to perform reliable research.
What you need
- A PhD or equivalent research experience in a quantitative field such as machine learning, computer science, mathematics, physics, or engineering is required.
- Demonstrated ability to formulate original research questions and successfully execute complex research projects from hypothesis to publication.
- A strong understanding of modern Large Language Models (LLMs) and the current frontier AI research landscape is necessary.
- Excellent experimental design skills, quantitative reasoning, and sound scientific judgment are essential for this role.
- The candidate must possess the ability to rapidly assimilate unfamiliar technical literature and quickly develop expertise in new research areas.
- Strong proficiency in Python is required to independently build necessary research prototypes and evaluation pipelines.
- Excellent technical writing and communication skills are necessary for documenting research findings.
Nice to have
- Strong publication record
Benefits
- $250,000 to $350,000 salary
Skills and tech stack
Questions candidates ask
What is the remote policy for this role?
The posting does not specify the remote work policy for this position.
What is the compensation range for this role?
The salary for this position ranges from $250,000 to $350,000.
What is the technology stack used in this research?
The required technical skill mentioned is strong Python for building research prototypes and evaluation pipelines.
What is the focus of the research team?
The research is deliberately focused on frontier STEM evaluation, synthetic data, hallucination and reliability, and agentic science.