Titan Title Review AI
Real-estate title review

Case study

AI Systems / RAG

Prototype

Dhaka, Bangladesh

AI Systems / RAG · Detailed breakdown

Project overview

Why this build
mattered.

Turns real-estate title PDFs into a cited, eight-section ALTA Title Review Summary, and learns from the edits the operator makes.

Snapshot

Real-estate title review / Prototype / Python, BAML, Qdrant

PythonBAMLQdrantBGE-M3Streamlit
Cited 8-section summary
output
BM25 + BGE-M3 + RRF
retrieval
Learns from edits
feedback

System views

Screens and
surfaces

What shipped inside

Key features
and modules

The architecture, workflow and results that made this project useful.

Highlight

PDF parsing with pdfplumber and Docling, structured extraction with BAML

Highlight

Hybrid retrieval: BM25 and BGE-M3 in Qdrant, fused with RRF and reranked

Highlight

Operator edits feed back into later summaries; Streamlit interface

Delivery summary

Build notes

Project essentials

Titan Title Review AI

Real-estate title review · Prototype · AI Systems / RAG

PDF parsing with pdfplumber and Docling, structured extraction with BAML1
Hybrid retrieval: BM25 and BGE-M3 in Qdrant, fused with RRF and reranked2
Operator edits feed back into later summaries; Streamlit interface3

Let's build something worth building.