Senior Clinical Informaticist

Mosai

1d•Hybrid

About The Position

Mosai is looking for a Senior Clinical Informaticist to be the clinical conscience of its AI and help drive AI-assisted health care at home. Working with data scientists and ML engineers, this person will own the clinical ground truth used to train and evaluate models, including what gets labeled, how it gets labeled, and whether outputs are safe and correct for nurses, therapists, and hospice teams to use. As the first hire in the clinical informatics function, this hands-on senior leader will create annotation guidelines, label and adjudicate data, build gold-standard datasets, and lead clinical evaluation of models and LLM-based features. The role brings knowledge of how care is delivered and documented at home into the data and logic used by Mosai's AI, while helping build the clinical informatics team.

Requirements

Active, unrestricted clinical license as an RN (BSN or equivalent), Physical Therapist, Occupational Therapist, or Speech-Language Pathologist.
At least five years of clinical experience, including meaningful time in home health and/or hospice.
At least four years in clinical informatics, such as informatics nursing, clinical analysis, clinical data abstraction, or quality, aligned with recognized standards such as the ANA Nursing Informatics Scope and Standards of Practice or AMIA health informatics core competencies.
Core informatics competencies in clinical workflow analysis, data governance and quality, the system lifecycle (requirements, testing, validation, and go-live), and clinical decision support or quality measure logic.
Working understanding of how traditional supervised and unsupervised machine learning and LLM-based models are trained and refined.
Hands-on OASIS expertise, including the current OASIS-E data set and guidance manual.
Working knowledge of ICD-10-CM coding and home health payment and quality programs, including PDGM, HHVBP, and Home Health Quality Reporting.
Experience with clinical data labeling, annotation, or chart abstraction, ideally supporting machine learning or analytics teams.
Comfort with clinical terminologies, including SNOMED CT, LOINC, and RxNorm.
Working knowledge of HIPAA and PHI de-identification methods, including Safe Harbor and Expert Determination.
Precise, consistent judgment and ability to explain clinical reasoning clearly to non-clinicians.

Nice To Haves

Informatics certification, such as ANCC Informatics Nursing (NI-BC, formerly RN-BC), HIMSS CPHIMS or CAHIMS, AMIA ACHIP, or AHIMA CHDA.
Master's degree in nursing informatics, health informatics, or data science.
Home health certification, such as COS-C, HCS-D, or HCS-O.
Experience evaluating LLM output, building evaluation rubrics, or conducting clinical red teaming.
Familiarity with responsible AI practices in health care, including bias and fairness testing, human-in-the-loop review, and model documentation.
Experience with annotation tools, such as Label Studio, Prodigy, or Labelbox, and basic SQL or Python.
Experience with HomeCare HomeBase or another home health or hospice EHR.
Familiarity with hospice documentation and quality reporting, including HOPE, HQRP, and CAHPS Hospice.

Responsibilities

Own clinical annotation: design labeling schemas and versioned guidelines for clinical notes, OASIS items, diagnoses, medications, wounds, functional status, and other home health and hospice concepts.
Label and adjudicate data directly; engage appropriate in-house clinical resources, resolve annotator disagreements, make final clinical calls on edge cases, and build and maintain core gold-standard datasets for data science models.
Work with MLOps and data science teams to report and monitor model performance and respond to drift in key accuracy metrics.
Measure inter-annotator agreement using standard metrics, such as Cohen's or Fleiss' kappa; run calibration sessions, audit samples, and improve annotation guidelines and workflows.
Evaluate model predictions and LLM-generated text for accuracy, completeness, hallucination, bias, and clinical safety.
Design evaluation rubrics and help define clinical criteria for release of traditional ML and LLM-based models.
Translate home health and hospice workflows, including start of care, recertification, resumption of care, discharge, hospice admission, and IDG, into structured data, rules, and features.
Map clinical concepts to ICD-10-CM, SNOMED CT, LOINC, RxNorm, OASIS, and hospice HOPE data elements; support EHR integration, including HomeCare HomeBase, using HL7 and FHIR.
Define clinical data quality checks, document provenance and dataset versions, and apply HIPAA, minimum-necessary, and de-identification practices to labeling work.
Support development of AI capabilities by demonstrating how skilled clinicians perform the same tasks.
Partner with product and engineering on clinical requirements, pre-release validation, post-release model monitoring, and clinician feedback loops.
Recruit, train, and mentor clinical annotators and junior informaticists as the program grows.