Mini-project
IT, Software & SaaS
Legal document classifier
Sort legal documents and court decisions by area of law.
The brief
A legal research start-up needs incoming documents sorted by area of law so the right team sees them. Build a classifier on a legal text benchmark and test how it handles Indian judgements. Good looks like: accuracy per area, examples of confusion, and a short note on using it with Indian documents.
What you will build
- Explore the benchmark tasks
- Build a TF-IDF baseline
- Fine-tune a legal or general language model
- Compare results
- Try it on Indian judgements and discuss the gaps
Free, open data to build it with
- LexGLUE legal benchmark ↗ Hugging Face
- Indian court judgements ↗ OpenNyAI (Hugging Face)
Check each dataset's licence before you publish your work. Your mentor gives your team its own version of the brief (data, scope or region), so every team's project is its own.