Nov 29, 2025
MUHAMMAD GHIFARY
In the rapidly evolving landscape of AI, healthcare is frontier where the promise of AI meets real-world impact. Enter MedGemma, a model family developed by Google DeepMind, designed specifically to understand and interpret both medical text and images. Built on the robust foundation of the Gemma 3 architecture (Gemma team, 2025), MedGemma represents a significant step toward giving developers and clinicians the tools to build healthcare-oriented AI applications.
Perhaps most importantly, MedGemma opens the door for accessible, locally-deployable health AI systems. With open-model availability and modular design, developers can run it on private infrastructure, maintain control of sensitive health data, and build customized workflows for triage, patient intake, or image-based diagnostics — rather than relying purely on cloud-based black-box models.
In this article, I’ll walk you through how we can build a simple AI healthcare assistant that runs entirely on our own local machine — thanks to MedGemma’s accessibility and domain-specialized power.
The personal AI doctor I built is just a simple web-app prototype featuring a conversational UI that supports both chat and voice commands. It runs entirely on a local machine without any Internet connection. Users can also upload images alongside particular instructions for the virtual doctor to perform medical image analysis. The system can respond to user prompts with text and, optionally, speech output, as illustrated in Figure 1.
Figure 1: Personal AI Doctor Demo
The prototype is fully written in Python, comprising frontend and backend services. The frontend implements the user interface components, while all relevant AI models will be served in backend services, comprising:
MedGemma is a collection of open-model variants developed by Google DeepMind specifically optimized for medical-domain tasks involving both text and images. Built on the underlying architecture of the Gemma 3 family, it brings healthcare-focused capabilities into a freely accessible model ecosystem.
MedGemma currently comes in two variants: a 4B-parameter multimodal version (capable of ingesting medical images + text) and a 27B-parameter text-only (and/or multimodal) version. The multimodal variant uses an image encoder, SigLIP (Zhai et al., 2023), which was pre-trained on a large corpus of de-identified medical imager (e.g., chest X-rays, dermatology photos, pathology pictures).
Because MedGemma inherits the capabilities of Gemma 3 (multimodality, long context, efficient architecture), developers can build healthcare applications that have strong baseline performance and benefit from the engineering work behind Gemma 3. Therefore, to further understand MedGemma, it’s worth understanding Gemma 3’s architecture in more details.
Gemma 3 is the base models for MedGemma sized from ~1B parameters up to ~27B parameters. Its design goals are as follows: