PI Discovery Engine
From research topics to investigator cards — discover collaborators through publication and expertise graphs.
The problem: Finding the right advisors for RA/PhD applications is a search and triage problem. Existing tools (Google Scholar, lab websites) require manual effort to compile and compare candidates.
The solution: PI Discovery automates the initial screening by searching Semantic Scholar and OpenAlex APIs, deduplicating and ranking candidates, and generating decision-ready PI cards — including publication history, h-index trends, country, and outreach status.
How It Works
Contact Pipeline Stages
FAQ
Why build a PI discovery tool?
Finding the right advisors for RA/PhD applications is a search and triage problem. Existing tools (Google Scholar, lab websites) require manual effort to compile and compare candidates. PI Discovery automates the initial screening.
What APIs does it use?
Semantic Scholar API for paper search and citation metrics, OpenAlex for author affiliation and country data. Both are free and openly accessible.
Is it usable for others?
The current local prototype supports topic and country filtering and ranked PI cards. A public release has not yet been prepared.
Tech Stack
Availability
Working local Python 3.11+ software prototype; public repository and setup instructions pending.
Key Takeaways
- Searches Semantic Scholar and OpenAlex APIs to find PIs working on specific research topics, with deduplication and relevance ranking.
- Generates decision-ready PI 'cards' including publication history, h-index trends, recent funding, country, and email contact status tracking.
- Streamlit web app with interactive filtering by country, topic, and minimum publications.
- Contact pipeline tracks outreach status across 7 stages: not_started → reading_papers → drafting_email → sent → replied → meeting_scheduled → declined.