Research interests

Research interests tell Atom what you work on. They are the single biggest factor in how well your grant recommendations match, and they belong to a project: each project has its own set, and the project you are working in decides which grants reach your dashboard and your inbox.

Where to edit them

Select Projects in the sidebar, open a project, and edit its Research interests section. Each line is one interest, and each line saves on its own as soon as you leave it, so matching improves with every refinement rather than waiting on a final save.

The project's lead PI and its admins can edit the interests, and so can an administrator at your institution, which lets a research development office write interests for the researchers they support.

Three ways to build them

• By hand. Type an interest per line. Keywords, a sentence, or a full abstract pasted from a recent paper all work.
• Generate with Atom AI. Atom reads the project name and the lead PI's background and drafts a set of interests, each pairing a method with what it is applied to. They are fully editable, so treat them as a starting point.
• Upload a document. Add a CV, research statement, or biosketch as a PDF or Word file, and Atom pulls interests out of it. Uploaded interests are merged with what is already on the project, so re-uploading does not create duplicates.

You can combine all three, and upload stays available whenever you can edit the project, so you can add a new document's interests long after setup.

Writing interests that match well

Atom uses semantic search to compare your interests to the full text of every opportunity, so you match on meaning rather than exact keywords. A grant about "machine learning for drug discovery" can match an interest in "computational methods in pharmacology."

Specificity is what makes that work:

• Pair a method with what it is applied to. "Ultrafast X-ray spectroscopy for quantum materials" matches far better than "materials science."
• Cover each distinct thread. A few precise lines beat one broad one.
• Split when they drift. If a project's interests no longer sit together, that is usually a sign the work has become two projects.