Łukasz Gandecki

Łukasz Gandecki is the founder and technical lead of TheBrain Software House and the creator of BookGenius, an interactive reading platform that adds character context, music and semantic search without distracting from the book itself. His career spans open-source JavaScript tooling, enterprise GraphQL, real-time infrastructure and AI-assisted product development.

Before BookGenius, Gandecki co-founded xolv.io, helping enterprise customers implement GraphQL, and worked across the Meteor and Apollo ecosystems. A project for 2K Games led him into real-time gaming data, DynamoDB and infrastructure tooling. He also developed the cypress-cucumber-preprocessor, which brought Gherkin-style specifications to Cypress testing, before transferring the project to another maintainer in 2022. His wait-for-expect package helps developers test asynchronous behavior without arbitrary delays.

BookGenius originated when Gandecki struggled to follow unfamiliar figures in a book about an American presidential campaign and built an AI companion that supplied contextual summaries and images. The experiment expanded into immersive fiction, including adaptations of The Snow Queen and 1984, which he demonstrated at AI Engineer World’s Fair 2025.

• Spoiler-aware interactive reading: BookGenius answers questions using only material the reader has already encountered. Character avatars, narration and scene-responsive music follow the current page; its open-source book player makes the interactive architecture available to other developers.

• Semantic search for remembered scenes: Embeddings let readers locate passages by describing what happened instead of recalling exact wording, then return to their original place. Longer queries can analyze previously read material while maintaining the same spoiler boundary.

• Human-directed generative experiences: Language models identify characters, scenes and emotional tone, while structured metadata coordinates illustrations and generated soundtracks. Human editors judge whether those elements actually suit the story.

• Pragmatic AI-assisted prototyping: Gandecki builds disposable early versions quickly, adds tests as a product stabilizes, and prioritizes refactoring according to code quality, real-world friction and ease of improvement.

More recently, he rebuilt BookGenius around Convex, favoring reactive TypeScript backends, end-to-end types and infrastructure understandable to coding agents, while retaining Cloudflare R2 for asset delivery. His Sessions recording project applies similar attention to developer experience, indexing macOS recordings by application, window and file while letting users exclude applications from capture.