Talent Graph Engine — 100K+ STEM Profiles, One Connected View

Researchers and engineers don’t live in a single database. Their signal is spread across conference proceedings, university pages, patent filings, GitHub profiles, lab websites, and dozens of other public sources — none of which talk to each other. A recruiter trying to understand who works on what, who trained under whom, or which lab is producing strong candidates in a specific subfield has to piece it together manually.

The relationships matter as much as the individuals. Knowing that a candidate co-authored with a well-published researcher, or moved from a research lab to a startup, reveals quality and trajectory that a keyword search on a CV never will. But without a connected view, those patterns stay invisible.

Scraped and structured data from 500+ public sources into a unified graph of 100K+ STEM and academic profiles. Each node — person, lab, company, institution — is linked by co-authorship, affiliation, advisory relationships, and career transitions. A graph UI lets users visually explore research and industrial connections across people, groups, labs, and companies.

In-degree and out-degree analysis surfaces hidden influencers and talent clusters. Shortest-path, minimum spanning tree, and other graph algorithms identify specific skills and talent concentrations across the network. A RAG integration resolves unstructured natural language queries — ask “who are the strongest mechanistic interpretability researchers who’ve published at ICML and have industry experience?” and get ranked results with full end-to-end profile data: education, work history, skills, achievements, awards, and a SWOT analysis per profile.