RareGPT
Innovating Rare Disease AI: Proving Concept with Custom GPT Development
Type
Case Study
Timeframe
4 weeks
Toolkit
HTML, JS, LLMs
Visit site
Unavailable
Challenge
Creating a custom GPT for rare diseases involved sourcing high-quality, expert-validated data, which was both time-consuming and required deep domain knowledge. The initial phase had to be conducted internally to prove viability, adding the challenge of limited resources and expertise in AI development.
Solution
By leveraging internal resources and my digital design skills, I developed and trained the model on a local server, proving the concept without external help or additional costs. This endeavor not only enhanced our understanding of AI training but also shifted our team's focus from abstract AI discussions to practical applications in the rare disease community, demonstrating a tangible benefit and readiness for further development.
Proved AI concept internally, enhancing our technical skills and moving from abstract ideas to practical rare disease solutions.
After briefing, I gathered user insights from a group of relevant user types, and began creating a complete front and back-end UX audit. I broke down the existing platform to identify the nessessary features and components, highlighting problem areas with a priotity traffic light system. I was then able to propose suggestions for improvements and additional features.
