Symbrex

• 

  Model pipeline

Inspiration

Both me and Rohan studied biomedical engineering and were personally involved with NHS. We saw how inefficient, centralised and outdated the current supply chain is, which inspired us to make it better.

What it does

Symbrex estimates the risk of the hospital equipment suppliers based on usage and satellite data. In case of predicted disruption, we provide alternative suppliers for the same item.

How we built it

We used data acquired from UCLH. For most of the LLM/VLM tasks, we used Claude Sonnet 3.5. Satellite imagery was acquired using coordinate scraping and google-earth-engine API.

Challenges we ran into

Low quality satellite images proved to be tricky to analyse. For now, we solved the problem by using google-earth-engine, but it's only a temporal solution, as higher quality providers are needed.

What's next for Symbrex

Incorporating more inputs would be natural next step. We're planning to introduce more robust market sentiment analysis that would scrape and analyse company financials. Additionally, it would be useful to gather more clinical setting data.

Built With

• claude
• flask
• google-earth-engine
• python
• react