Skip to content

Project

Contract and public tender analysis

Role
AI Engineer
Context
Global Hitss
Stack
  • python
  • amazon bedrock
  • rag
  • embeddings
  • aws

o200k_base · 132 tokens · what the model reads

Generative AI and RAG applied to public tenders and contracts.

The problem

Public tender notices and government contracts are long, dense documents written in their own legal register. Finding one specific clause (a deadline, a penalty, an eligibility rule) means reading dozens of pages per document.

The hard part is not generating text: it is making sure the answer came from the right document and can be checked. In a legal context, an invented clause is not a cosmetic error.

The solution

Retrieval augmented generation over the document base, with every answer anchored to a quotable passage from its source document.

Back to projects