LeapSpace

Zochi is the first AI system capable of autonomously completing the entire scientific research process, from hypothesis generation to peer-reviewed publication, producing state-of-the-art results. Unlike prior systems limited to narrow, predefined tasks, Zochi excels in addressing research challenges at the forefront of artificial intelligence. Its effectiveness is validated by multiple peer-reviewed publications accepted at ICLR 2025 workshops, underscoring Zochi's ability to generate novel and academically rigorous contributions. Zochi identified a critical bottleneck in AI development: cross-skill interference in parameter-efficient fine-tuning. When adapting models to multiple tasks simultaneously, improvements in one skill often degrade others. To address this, Zochi developed CS-ReFT (Compositional Subspace Representation Fine-tuning), focusing on representation editing rather than weight modifications.

scienceOS is an AI-powered research platform built to accelerate scientific literature workflows by giving researchers fast, reliable access to a massive database, more than 225 million research papers via a chat-based interface. The core “AI science chat” lets you ask questions, get answers grounded in published literature, and even generate tables or diagrams summarizing findings. If you upload PDFs, the “multi-PDF chat” can parse up to eight documents per session and extract key passages, figures, and tables to help you digest papers quickly; it can also generate structured summaries of papers (e.g., intro, methods, conclusions), highlighting main findings, limitations, and key data. Alongside that, scienceOS includes an AI reference manager; you can store and organize up to 4,000 PDFs or citations in a personal or shared library, import external references (e.g., from Zotero), and chat with your own collection, useful for drafting literature reviews and building bibliographies.