Find papers, people, institutions, communities and opportunities from one place.
Scan by entity type, then open the next useful thread.
The strongest cross-entity jumps appear after the first useful query.
We combine publication metadata, author neighborhoods and dataset signals to improve retrieval quality for computational biology teams working across papers, code and experiments.
We study how follow edges, review quality, graph proximity and freshness can be blended into an explainable feed optimized for high-signal research discovery.
We evaluate rubric design for structured reviews, moderation queues and reviewer calibration in technical communities.
We propose ledger-based moderation records that improve accountability, appeals and policy learning in research products.
We curate paper, code and dataset benchmarks for biology-focused language models under provenance and reproducibility constraints.
We precompute neighborhoods over works, authors, communities and institutions to speed exploratory search.
This paper combines institution verification, topical expertise and trust snapshots to route methodological questions to the right specialists.
We align code repositories, datasets and papers with provenance-aware linking that survives version churn and renamed assets.
We extract actionable collaboration intent from follows, saves, reviews and topic overlap without collapsing trust into a single opaque metric.
We present a graph-native product architecture for scholarly sensemaking where work pages, topic maps and researcher trust signals reduce time-to-understanding for new literature.
This work formalizes a claim workflow that separates platform users, researcher profiles and author entities, improving identity resolution without degrading product usability.
We show that persistent graph neighborhoods and operator-facing retrieval memory improve literature navigation across papers, people and opportunities.
We study open-weight assistants that combine paper summaries, dataset cards and experiment context for biology teams.
We compare dense and sparse retrieval strategies for scholarly search when operators need interpretable reasons and controllable ranking.
We trace dataset reuse across scholarly graphs to improve auditability, discovery and replication planning.
We use graph signals and topical expertise to connect researchers with fellowships, residencies and collaboration openings.
Lab-based community around research infrastructure, graph-native discovery and evaluation systems.
InstitutionETH Zurich
High-signal discussion for scholarly graphs, entity resolution and research discovery systems.
Institution-linked group for retrieval pipelines, literature tooling and multimodal discovery infrastructure.
InstitutionMassachusetts Institute of Technology
Invite-only calibration circle for structured reviews, trust signals and moderation quality.
Industry-linked research partner focused on graph-native scholarly tooling, retrieval systems and human-in-the-loop discovery workflows.
Infrastructure partner for literature-scale indexing, provenance and research workflow operations.
Foundation partner backing calibrated review programs, trust research and reproducibility-centered evaluation layers.
Build literature-aware retrieval and provenance infrastructure for scientific datasets, code links and multimodal paper navigation.
Lead a graph-native discovery program spanning entity resolution, retrieval quality and trust-aware product instrumentation.
Run calibration cohorts for structured reviews, moderation quality and trust measurement in networked science products.
