Natural Language Measurement Specialist

Natural Language Specialist, Educational Measurement & AI   College Board –   Learning & Assessment   Location:     This is a remote role. Candidates who live near CB offices have the   option   of being fully remote or hybrid (Tuesday and Wednesday in office). All CB employees   are required to   occasionally travel to meet in person for business purposes.   Role Type :    This is a full-time position   About the Team     The Automated Scoring team provides critical insights and tools to support the design, delivery, and continuous improvement of digital assessments. We  operate  at the intersection of educational measurement, data science, and emerging AI technologies. Our work spans the measurement of language-based constructs, the development of large language model (LLM) systems for feedback generation and annotation, and research to ensure the validity, fairness, and reliability of our systems.    We are a collaborative, mission-driven team that values both psychometric rigor and technical  expertise . We combine modern machine learning approaches, including large language models, with strong measurement principles to create scalable, trustworthy solutions that expand opportunity for students.    About the Opportunity      As a Natural Language Specialist, you will help define and advance how language-based performance is measured in high-stakes educational settings. This role sits at the intersection of natural language processing, large language models, and educational measurement, and is ideal for someone who  pairs  strong measurement training with a working knowledge of modern AI.    You will translate measurement constructs into LLM-based feedback and annotation systems, design and conduct the studies that  establish  their validity, fairness, and reliability, and ensure that what our models produce holds up to rigorous psychometric standards.   You will s erve as the measurement authority for cross-functional partners—psychometricians, engineers, and data scientists—shaping how language-based constructs are defined, evaluated, and applied across the   team's portfolio of systems .   Y our primary lens will be measurement: defining what feedback and annotations mean,  evidencing  that they mean it, and improving them over time.    In this role, you will:     Natural Language Measurement & Psychometrics  (40%)     Define and operationalize language-based constructs for automated annotation and feedback generation    Apply psychometric principles—reliability, validity, dimensionality, and measurement invariance—to LLM-based feedback and annotation systems    Design and lead validity studies, including human–machine agreement, rater comparison, and fairness analyses across subgroups    Develop and apply methods for detecting and mitigating bias in language-based scores    Establish /Recommend    annotation guidelines, feedback quality criteria, and standards for acceptable model performance    Translate measurement requirements into specifications that guide model development and evaluation    LLM & AI Development  (30%)     Contribute to prompt design, fine-tuning, and evaluation of LLM-based feedback and annotation systems    Develop and refine machine learning models for measuring language-based constructs    Build evaluation frameworks that connect model behavior to measurement outcomes    Collaborate with senior team members to translate measurement findings into production systems    Implement high-quality, maintainable code for model development and evaluation    Research & Validation  (10%)     Lead and contribute to research studies that evaluate model performance and support assessment validity    Apply statistical and psychometric methods to analyze results and inform model improvements    Document methodologies and findings in a clear and rigorous manner    Stay current with advances in educational measurement, NLP, and learning science    Data Engineering & Pipelines  (10%)     Prepare and curate datasets for measurement studies and model evaluation    Support reproducible data processing workflows for training, evaluation, and monitoring    Partner with engineers to integrate feedback and annotation models into scalable systems    Team Operations & Collaboration  (10%)     Collaborate closely with psychometricians, data scientists, and engineers    Contribute to documentation, methodological standards, and team best practices    Participate in peer reviews and knowledge sharing    Actively raise the measurement literacy of the broader team—mentoring junior ICs, providing technical feedback on colleagues' work, and building shared standards   About You     You bring strong measurement training and a genuine interest in how modern AI can be used to measure language-based performance. You are excited about applying psychometric rigor to large language models in high-stakes educational settings.  You have:     A  Master’s  or PhD (or near completion) in a quantitative field such as Psychometrics, Educational Measurement, Quantitative Psychology, Statistics, Data Science, or a related discipline (measurement-focused training strongly valued)    A solid foundation in measurement theory, including reliability, validity, and fairness; familiarity with IRT, generalizability theory, or related frameworks is highly desirable    Experience analyzing language or text data, with exposure to NLP or large language models    Programming skills in Python and familiarity with data science libraries (e.g., pandas, NumPy,  PyTorch )    Demonstrated ability to conduct rigorous research (e.g., thesis, publications, or applied research projects)    Strong skills in statistical analysis and experimental design    Familiarity with working with structured and unstructured data    Strong attention to detail and a commitment to producing high-quality, reproduc