CellDX AI Autopilot: Agent-Guided Training and Deployment of Pathology Classifiers

2026-05-11

AI Summary

This work addresses the significant barriers in developing computational pathology AI models—namely, the high cost of data acquisition, dependence on GPU resources, steep machine learning expertise requirements, and engineering complexity. To overcome these challenges, we present the first pathology-specific training platform designed for general-purpose AI agents, enabling end-to-end automation through natural language interaction that guides users from data curation and hyperparameter tuning to model deployment. By integrating domain-specialized pathology agent capabilities with large language model (LLM) agents, our approach eliminates the need for task-specific customization. We introduce an iterative pairwise hyperparameter search strategy that reduces tuning costs by over 30-fold compared to exhaustive search. Built upon 32,000 whole-slide images and a multiple instance learning framework, the platform supports four classification paradigms, substantially lowering reliance on ML specialists and accelerating both research and clinical translation.

Abstract

Training AI models for computational pathology currently requires access to expensive whole-slide-image datasets, GPU infrastructure, deep expertise in machine learning, and substantial engineering effort. We present CellDX AI Autopilot, a platform that lets users -- from pathologists with no ML background to ML practitioners running many parallel experiments -- train, evaluate, and deploy whole-slide image classifiers through natural language interaction with an AI agent. The platform provides a structured set of agent skills that guide the user through dataset curation, automated hyperparameter tuning, multi-strategy model comparison, and human-in-the-loop deployment, all on a pre-built dataset of over 32,000 cases and 66,000 H&E-stained whole-slide images with pre-extracted features. We describe the agent skill architecture, the underlying Multiple Instance Learning (MIL) training framework supporting four classification strategies, and an iterative pairwise hyperparameter search (grid or seeded random) that reduces tuning cost by over 30x compared to exhaustive search. CellDX AI Autopilot is, to our knowledge, the first system to expose pathology-specialized agent skills and a pathology-specialized training platform to general-purpose AI agents (e.g. any LLM-based agent runtime), delivering end-to-end automated model training without requiring the agent itself to be domain-specific. The platform addresses both the ML-expertise bottleneck that limits adoption in diagnostic pathology and the engineering bottleneck that limits how many experiments a researcher can run cost-effectively.

Problem

• computational pathology
• AI training bottleneck
• whole-slide image classification
• machine learning expertise
• engineering scalability

Innovation

• AI agent
• computational pathology
• Multiple Instance Learning
• hyperparameter optimization
• natural language interface