Making a deep body of knowledge easier to explore.

What the client brought

A substantial legacy library and the ambition to make it useful as a modern research product.

What the work revealed

A useful research experience depends on reliable source identity and continuity between finding, reading, and questioning.

The decision that shaped the system

Make original sources the foundation of the product. Search finds candidates; application code retrieves exact passages and citations. Model assistance helps interpret the inquiry without becoming the authority for the text.

What was built

The engagement connected product definition, database conversion, experience design, architecture, and native development. Dimas carried nearly all of that work, building macOS and iOS applications with local hybrid search and on-device AI integration on iOS.

Where judgment mattered

Two semantic-search candidates were rejected when evaluation failed to show the required improvement. The working baseline stayed active. A later hybrid implementation was a separate milestone; the case does not claim benchmark superiority.

Current stage

Native applications have been built. Extended source-linked inquiry, evidence capture, and reader-continuity workflows remain in development and evaluation. Adoption and commercial outcomes have not been established.

Connected responsibilities

• Product strategy
• Source recovery
• Native applications
• Retrieval architecture
• On-device AI
• Commercial planning