NimbleLabs: Accelerating Healthcare AI Development Through Agentic AI
Conference
Shimgekar, SR, Goyal, A, Vassef, S et al. (2025). NimbleLabs: Accelerating Healthcare AI Development Through Agentic AI
.(2025), 8327-8329. 10.1109/BigData66926.2025.11401464
Shimgekar, SR, Goyal, A, Vassef, S et al. (2025). NimbleLabs: Accelerating Healthcare AI Development Through Agentic AI
.(2025), 8327-8329. 10.1109/BigData66926.2025.11401464
Extracting meaningful information from unstructured medical data is a major challenge in healthcare analytics, requiring substantial time, computational resources, and specialized expertise. Agentic AI introduces new opportunities to automate and streamline these workflows. We present a multiagent architecture that democratizes medical data analysis for data scientists, medical researchers, and healthcare practitioners. The system enables users to: (i) obtain comprehensive insights through automated dataset analysis; (ii) automatically integrate additional knowledge from supporting files; and (iii) develop predictive models without extensive machine learning expertise. The architecture contains six specialized agents: (i) Type Identification Agent, which classifies data (structured/unstructured) and performs privacy-preserving anonymization; (ii) Feature Identification Agent, which extracts dataset features; (iii) Feature Enrichment Agent, which generates contextually relevant keyword vocabularies for each feature based on user intent; (iv) Additional File Integration Agent, which uses semantic and keyword-based extraction to incorporate supplementary information from PDF, Excel, and CSV files; (v) Input-Output Optimization Agent, which determines ideal input and output features for machine learning based on user intent; and (vi) Modeling Advisory Agent, which recommends suitable predictive models. We evaluate the system across multiple medical data modalities. For healthcare providers, research institutions, and health-tech companies, this workflow enables faster decisionmaking, reduced data-processing costs, improved regulatory compliance, and the ability to transform raw medical data into actionable insights.
