The problem
Useful knowledge is scattered across documents, spreadsheets, archives, and shared drives. Finding a file is only the beginning of answering a question.
Enterprise knowledge
KIQTT
A permission-aware search and analysis layer for organizational knowledge.
↳ Created with my teamAt a glance
KIQTT is an enterprise knowledge-search project for people who need to find information across institutional documents. Our team’s documented work connects document ingestion, OCR, indexing, and retrieval-augmented generation so that source material can support answers. Access permissions are part of the system design, alongside the handling of different document formats and processing tasks. The project demonstrates the connection between document preparation, retrieval, and an AI response rather than treating search as a standalone chat interface. The public overview describes the engineering approach; it does not publish retrieval-quality measurements or deployment coverage.
System flow
- 01Files & access rules
- 02OCR & retrieval
- 03Source-based answers
Created with my team
The projects in this portfolio are a shared effort. Product decisions, architecture, engineering, and delivery belong to the people who build together. My responsibilities vary from project to project.
Who it serves
Employees and analysts searching internal organizational documents.
What we build
Our team’s work connects document ingestion, OCR, semantic retrieval, reranking, and question answering. Spreadsheet analysis adds a structured-data path alongside document search.
Engineering decisions
Windows agents and network-share ingestion feed processing pipelines. Active Directory roles inform access boundaries; an LLM router separates document and tabular analysis.
Capabilities
Word/PDF parsing, scanned-document OCR, semantic search, reranking, spreadsheet processing, and role-aware retrieval design.
Context & limitations
Architecture goals and implementation work are described together. Search quality and permission enforcement require evaluation on the organization’s own data; planned KPIs are not presented as results.
Technical foundation