Diego Santana
Applied Research Engineer
Builds collaboration tooling and graph-backed ranking loops for scholarly platforms.
Open to collaborateInstitution profile
Massachusetts Institute of Technology
Cambridge, US. Institution profiles help you understand where researchers, outputs, topic strength and operational trust gather in one place.
Institution snapshot
At a glance
24labs
181researchers
13works
4topics
Operator path
Institution operations and trust
Admin access appears only for verified operators and bootstrap-eligible primary affiliates.
Institution graph
See the institution as a connected system
Inspect nearby researchers, outputs, topics and circles without leaving the institution page.
Drag to move the map, use wheel or controls to zoom.Selected institution: Massachusetts Institute of Technology
institutionsignal 104 adjacent
Massachusetts Institute of Technology
Cambridge, US
Researchers
4 connected scholar(s)
Noah Darzi
Staff Research Scientist
Studies expert routing, retrieval systems and opportunity matching in scientific networks.
Open to collaborateOwen Chen
Postdoctoral Researcher
Works on literature-scale retrieval and scientific representation learning.
Open to collaborateSunwoo Kim
Industry Research Scientist
Bridges dataset-centric ML with reproducibility tooling.
Heads-down modeOutputs
13 representative work(s)
Literature-Aware Retrieval Pipelines for Multimodal Computational Biology
We combine publication metadata, author neighborhoods and dataset signals to improve retrieval quality for computational biology teams working across papers, code and experiments.
Multimodal Benchmark Curation for Biology LLMs
We curate paper, code and dataset benchmarks for biology-focused language models under provenance and reproducibility constraints.
Dataset Provenance Across Paper Graphs: Linking Results, Repositories and Reuse
We trace dataset reuse across scholarly graphs to improve auditability, discovery and replication planning.
Graph Neighborhoods for Scholarly Search and Topic Expansion
We precompute neighborhoods over works, authors, communities and institutions to speed exploratory search.
Probabilistic Claim Resolution for Author Entities in Scholarly Graphs
This work formalizes a claim workflow that separates platform users, researcher profiles and author entities, improving identity resolution without degrading product usability.
Reproducible Paper-to-Code Linking for Research Infrastructure
We align code repositories, datasets and papers with provenance-aware linking that survives version churn and renamed assets.
Verified Expert Routing with Institutional Signals in Research Networks
This paper combines institution verification, topical expertise and trust snapshots to route methodological questions to the right specialists.
Open-Weight Biology Assistants for Dataset-Centric Literature Triage
We study open-weight assistants that combine paper summaries, dataset cards and experiment context for biology teams.
Graph Memory for Literature Agents: Persistent Neighborhoods for Scientific Search
We show that persistent graph neighborhoods and operator-facing retrieval memory improve literature navigation across papers, people and opportunities.
Sparse Embeddings for Scholarly Discovery with Human-Readable Signals
We compare dense and sparse retrieval strategies for scholarly search when operators need interpretable reasons and controllable ranking.
Collaboration Intent Signals in Research Labs
We extract actionable collaboration intent from follows, saves, reviews and topic overlap without collapsing trust into a single opaque metric.
Trust-Aware Feed Ranking for Scholarly Collaboration Networks
We study how follow edges, review quality, graph proximity and freshness can be blended into an explainable feed optimized for high-signal research discovery.
Topic-Aware Opportunity Matching for Research Careers
We use graph signals and topical expertise to connect researchers with fellowships, residencies and collaboration openings.