Data Scientist - Research AI

- Design multimodal foundational model architectures that integrate data from imaging, text, waveforms, structured data, graph representations, and temporal embeddings. - Develop fusion, alignment, and cross-modal reasoning techniques, including early fusion, late fusion, and hybrid models. - Outline and execute methodologies for grounded clinical reasoning, retrieval-augmented inference, and graph-augmented attention. - Set up protocols for model lifecycle governance, including safe update cycles and drift-aware re-training. - Train extensive deep learning models, encompassing multimodal frameworks and domain-specific transformer systems, using real clinical datasets. - Fine-tune large language models (LLMs) for clinical tasks like reasoning, summarization, and instruction-following. - Construct retrieval-augmented pipelines utilizing embeddings and vector storage with clinically relevant context. - Devise evaluation metrics for reasoning quality, prediction accuracy, and model robustness. - Create reference-grounded training datasets and multimodal benchmarks for performance assessment. - Engage in practical experimentation with optimization strategies and model acceleration techniques. - Implement uncertainty modeling and clinically significant risk thresholds. - Generate interpretable reasoning pathways and attribution maps for various modalities. - Collaborate closely with the Representation team to ensure model alignment. - Partner with clinical subject matter experts to integrate domain reasoning into reinforcement learning or rule-guided approaches. - Provide leadership and mentorship to scientists and engineers, building out a specialized modeling team.