PorePatrol AI

• 

  Screenshot of demo

• 

  Logo

Inspiration

We believe that the biggest challenges in society arise from the bottleneck of human expertise. Be it access to educators, social workers, or medical professionals.

Our vision for the next era of AI is one in which everyone has limitless access to professional help and guidance — especially in healthcare.

We’ve built a multi-modal GenAI product for therapeutic dermatology to illustrate both the potential and versatility of this technology.

What it does

Health anxiety causes major distress and costs the NHS 400m annually. (And who’s more anxious than pimply teens?)

We’re launching an AI health assistant that mitigates anxiety about skin conditions like acne by providing personalised guidance, ongoing symptom monitoring, and ever-available reassurance.

But it doesn’t stop there. Building on widely-capable LLMs, we’re able to rapidly branch out to other forms of health guidance without changing the tech stack — bringing a user’s existing data along as context.

How we built it

We spin up a custom instance of the multimodal LLaVa model to medically interpret user images. With 4-bit quantisation, LLaVa can be run on-device for further image privacy.

Claude then integrates image descriptions with user history, harnessing its extended context and broad knowledge for tailored therapeutic suggestions.

Users can then converse with our agent, discussing their concerns and receiving compassionate, but well informed guidance

Built With

• amazon-ec2
• claude
• gradio
• llava
• python
• pytorch