University of Maryland School of Medicine
The AI Grant Matcher
finds the funding you'd never have time to look for.
Every day, the Grant Matcher scans federal and foundation funding announcements and matches each new opportunity against the research profiles of SOM faculty — so relevant funding finds you, instead of the other way around.
Why it exists
Three goals
Grants.gov alone posts thousands of notices a year. No investigator can watch it all — and the right opportunity is easy to miss in a week of clinic, teaching, and lab work. The Grant Matcher was built to carry that watch for you.
Never miss a fit
Every new NIH, CDC, HRSA, DoD, NSF, and foundation announcement is compared against your research profile the day it posts.
Respect your time
You get one short email — only when something matches you — with the deadline, the agency, and why it matched. Nothing to search, nothing to configure.
Get smarter with use
Every piece of feedback you give teaches the matcher what a good match looks like for you and for colleagues in your field.
What you'll receive
The weekly notification email
When the matcher finds opportunities for you, you'll receive a personal AI Grant Match Notification each Tuesday — a roundup of the past week's new opportunities that matched your profile. Each grant shows the deadline, your confidence score, the evidence behind the match, and one-click Good match / Not relevant feedback buttons. Near-identical announcements (like one NOFO posted per country) are grouped into a single entry.
The Excel report
The complete week's detail also exists as a workbook — a Summary tab listing every grant, then one tab per grant with each matched faculty member's department, email, confidence, match type, and match evidence, plus direct links to the grant page and faculty profiles. Department research administrators and the project team use it for the full picture; ask if you'd like it included with your email.
How to read a match
- 62%Confidence — how strongly your profile aligns with the grant. Only matches at 50% or higher are sent; the strongest run into the 90s.
- keyword / AI / keyword + AIMatch type — keyword means terms from your profile appear in the grant text; AI match means the system recognized a conceptual fit even without shared wording; keyword + AI means both agree — the strongest signal.
- ≈Evidence terms — for AI matches, the parts of your own profile and publications that most resemble the grant, so you can see why it surfaced.
- SeriesWhen an agency posts many near-identical announcements (for example, one per country), they're grouped into a single card so your inbox isn't flooded.
Under the hood, briefly
How matches are made
Your match profile starts with the Research/Clinical Keywords on your SOM faculty profile, then is enriched automatically with your public research footprint — publication titles from PubMed and ORCID, active NIH awards, and registered clinical trials. Each new grant is compared against that profile two ways at once: exact keyword matching for precision, and AI semantic matching to catch conceptual fits your keywords don't spell out. Grants UMSOM isn't eligible for — and announcements that aren't really research opportunities — are filtered out before anything reaches you.
The part only you can do
Help train the matcher
The matcher learns from two things: what your profile says, and what you tell it about its matches. Three ways to make your matches better — in order of impact:
Refine your faculty-profile keywords Biggest lever
Your Research/Clinical Keywords are the foundation of every match. Specific beats broad: "pediatric traumatic brain injury" will match far better than "pediatrics" and "brain." Five minutes updating your SOM profile improves every future match.
Click 👍 or 👎 on every match One click
Each match in your email carries Good match / Not relevant buttons. One click records your verdict — no login, no form-hunting. If a whole email misses, use the single "none of these are relevant" link at the bottom. These clicks are how the matching thresholds get tuned.
Tell us why Most valuable detail
After a 👍/👎 click you'll see an optional comment box. A sentence like "this requires running a patient registry — I'm a bench scientist" teaches the system a distinction it can apply across the whole school.
Common questions
FAQ
I got a match outside my field. Why?
Usually one of two reasons: a broad keyword on your profile (see lever #1 above), or the AI recognizing a conceptual neighbor that isn't actually your work. Either way, a 👎 click is exactly the signal that fixes it — for you and for colleagues with similar profiles.
How often will I get email?
One email per week at most: a Tuesday roundup of the opportunities that matched you in the past week. No matches, no email.
Who sees my feedback?
The UMSOM Office of Information Services project team, in aggregate, to tune the matcher. Your clicks are never shown to other faculty and have no bearing on anything besides match quality.
Does a high confidence score mean I should apply?
It means your profile aligns strongly with the announcement's language — it is a lead, not an endorsement. Always read the full announcement for eligibility, scope, and fit, and loop in your department's research administration early.
How do I stop receiving these emails?
Every email carries an "Opt out of these emails" link in the footer — one click records your opt-out and you'll stop receiving notifications. You can re-enroll any time by contacting the Office of Information Services.
Is this the only place I should look for funding?
No. The Grant Matcher is designed to complement — not replace — your other funding-discovery channels: program officers, professional societies, and department announcements.
Questions or ideas?
The AI Grant Matcher is built and operated by the UMSOM Office of Information Services. For help, feedback, or to suggest a funding source we should monitor, contact help@som.umaryland.edu.
The information provided by this application is intended to notify SOM faculty of potential grant opportunities that may align with their expertise. It should be used alongside other methods of identifying funding and should not serve as the sole source for grant discovery.