Data Manager — Multimodal Medical Foundation Models

You will lead data operations for a cutting-edge research group developing 3D medical multimodal foundation modelsand agentic clinical AI systems. These models rely on extremely high-quality, well-structured, and compliant datasets—including 3D medical imaging volumes (MRI, CT, PET), clinical text corpora, annotations, and multimodal metadata.

Your job is to own the end-to-end data lifecycle: acquisition, ingestion, cleaning, versioning, labeling, quality control, governance, and delivery to researchers. You are the central node ensuring our foundation model teams and medical agent teams have clean, scalable, well-documented data pipelines.

This is a pivotal foundational role—without great data, large models cannot be great.

• Oversee ingestion and processing of 3D medical volumes (DICOM, NIfTI, MHA) and associated clinical texts.
• Build automated pipelines for metadata extraction, de-identification, slice/series validation, and cohort structuring.
• Manage large-scale internal datasets and external research datasets (BraTS, LiTS, MIMIC-CXR, CheXpert, MosMed, etc.).

• Implement scalable data storage, cataloging, and retrieval systems for multimodal training data.
• Own dataset version control, lineage tracking, reproducibility, and dataset documentation.
• Collaborate with ML systems engineers on high-throughput data loaders, sharding strategies, and caching mechanisms.

• Lead medical annotation workflows with radiologists, medical students, and labeling vendors.
• Create guidelines for ROI labeling, segmentation, captioning, report alignment, and case-level curation.
• Build semi-automated labeling pipelines using model-assisted tools.

• Enforce strict standards on data quality, completeness, consistency, and bias control.
• Ensure adherence to medical data privacy, HIPAA-equivalent frameworks, and institutional data-sharing rules.
• Manage PHI de-identification, audit logs, access control, and compliance approvals.

• Work closely with foundation-model researchers to understand data needs for model training.
• Partner with agentic system designers to supply structured datasets for clinical reasoning tasks.
• Collaborate with foundational engineers on data access layers, performance bottlenecks, and dataset optimization.

• The foundation model relies on high-quality 3D and textual data at scale.
• You shape the data pipelines enabling next-generation medical AI agents.
• You ensure clinical-grade governance, safety, reproducibility, and trust.
• Your systems become the backbone for research, experiments, and deployments.

For candidates motivated by the intersection of data, healthcare, and machine learning, this is a high-impact opportunity.

• Strong experience managing large multimodal or imaging datasets, ideally medical imaging.
• Proficiency with DICOM/DICOMweb, NIfTI, PACS systems, and medical imaging toolkits (dicompyler, pydicom, MONAI, ITK).
• Experience with ETL pipelines, distributed data systems, and cloud/on-prem storage.
• Knowledge of metadata standards, ontologies, and text–image linking strategies.
• Comfortable working with Python, SQL, and data tooling (Airflow, Prefect, Dagster, DBT, Delta Lake, etc.).
• Understanding of data privacy, de-identification, and compliance requirements in healthcare.
• Strong communication skills and the ability to coordinate between engineers, researchers, clinicians, and data partners.

• Experience with vector databases, multimodal retrieval, or embedding store design.
• Familiarity with annotation tools (Labelbox, CVAT, iMerit, custom MONAI Label pipelines).
• Prior work with clinical NLP datasets or multilingual Indian medical corpora.
• Experience conducting bias audits, dataset characterization, or quality scoring at scale.
• Contributions to open datasets, benchmarks, or data documentation frameworks.