R.A.I.D.E.N.

WHAT IS PROJECT R.A.I.D.E.N.?

R.A.I.D.E.N. (Reasoning AI Doctors Enhanced Network) is an agentic AI system to make access to specialist medical expertise more efficient, while ensuring results are aligned thanks to the feedback of a human medical doctor. 

The main features of R.A.I.D.E.N. are:

• it simulates a team of specialist medical doctors that analyze symptoms and generate a possible diagnosis

• it is based on the Google Gemini 2.5 flash and flash-lite LLM, and the Google ADK framework for agentic AI

• a human medical doctor must review the indications from the AI specialists and can provide feedback before final approval

• it can produce both PDF and mp3 files for text and audio version of the medical report.

WHAT MOTIVATED R.A.I.D.E.N.?

Access to specialist medical doctors is often challenging, with long waiting time and high cost for both the patient and the healthcare system. At the same time, the General Practitioner (GP) of the patient is often involved in the referral and follow-up, making the whole procedure quite demanding, especially during crises. LLMs and agentic AI have the potential to help mitigate this challenge.

Project R.A.I.D.E.N. was submitted as a capstone project to the Agents Intensive course on Kaggle with Google.

HOW WAS R.A.I.D.E.N. DESIGNED?

First, the logo. It was made with Google Nano Banana!

Concerning the main architecture, it is based on Google Gemini and the Google ADK for agentic AI. A parallel agent simulating a team of specialist medical doctors (e.g., cardiologist, neurologist, gastroenterologist, CBT psychologist, and an internal medicine agent) receives the query with symptoms and provide a diagnosis. After that, there is a loop agent where a human doctor provide feedback on the diagnosis. If the human doctor approves it, a final report in both PDF and mp3 format is created.

A picture summarizing the architecture is below, together with a video showing how to use it. 

PROS, CONS, AND FUTURE WORK

✅ R.A.I.D.E.N. is a tool that leverage advanced LLMs and agentic AI to make access to medical expertise more efficient, with benefits for both patients and healthcare systems, while ensuring alignment thanks to the human doctor in-the-loop.

🆘 Automatic agentic systems can make mistakes, so it is important to ensure that the feedback loop works as expected and that hallucinations, if happening, are detected.

💡The next step is to add more specialist medical doctor agents, and provide them with tools to access information and perform specific tasks. For example, an agent could be fine-tuned to read x-ray. If the system is satisfactory, a first pilot could include testing R.A.I.D.E.N. in a medical setting to collect feedback from real medical practitioners.

If you have ideas, comments, suggestions, feel free to get in touch!

RESOURCES

• Project code (Kaggle Notebook) ▶️here

• The Kaggle Writeup of the project ▶️here