All projects

CLIFF / Web · Academic research

Making the reasoning visible.

An exploration of multiple AI perspectives for medical image analysis, built as a semester project.

CLIFF
CLIFF illustrative analysis interface comparing sample model outputs
CLIFF — info

CLIFF

Design & development
by Siddhartha

Built with

  • Next.js
  • React
  • Python
  • FastAPI
  • Supabase
  • OpenRouter

Why I made it

For my fifth-semester major project, I wanted to explore how several AI models could examine the same image, and how their individual reasoning could be made easier to compare.

CLIFF
CLIFF local demo history with fictional sample records
Keeping a history

What it does

The Clinical Learning and Findings Framework sends an image to several vision models, brings their candidate findings together, and shows the individual outputs alongside an aggregated result. Analysis history keeps the process traceable.

CLIFF
CLIFF local demo detail with an illustrative comparison record
Returning to a result

The details I cared about

I designed the web interface around a clear upload-to-analysis flow, readable results, and access to the reasoning behind them. It is an academic exploration, not a clinically validated diagnostic tool.

CLIFF
CLIFF model directory listing the models used by the research prototype
The model directory
Another thing I made

Superplann