Keep your CV current with Claude Cowork.

Cowork works on files on your own computer. So the first thing is to put your CV where it can reach it, and tell it which folder that is.

1. Make a folder for it. One folder, call it CV, with your current CV inside. A Word file is easiest, but it can read a PDF too.
2. Give Cowork access to that folder. When you ask it to work on the CV, it will request the folder. Point it at the one you just made.
3. Tell it to work on a copy. Say "keep my original and make the changes in a new version." Now the file you have had for years is never the one being edited.

This is the part that saves the most time and also the part to be most careful with. Cowork can search PubMed for papers under your name and list the ones not already in your CV. The catch is that PubMed does not know which "J Smith" is you.

So you make it match on more than the name. Paste this in, with your details filled in:

Search PubMed for publications by [FULL NAME, e.g. Pectasides E], [SPECIALTY] at [INSTITUTION]. My ORCID is [ORCID, or "none"].

- Match on my name together with my affiliation or ORCID. Do not include papers by other authors who share my name.
- Compare against the publications already listed in my CV and show only the ones that are not already there.
- For each new one, give the full citation in [STYLE, e.g. AMA] and mark it [CHECK].
- Do not add anything to the CV yet. Show me the list first so I can confirm each one is mine.

Then you read the list before any of it goes in. One rule: never add a citation you have not confirmed is yours. Name collisions are common, and a wrong paper on a CV is worse than a missing one. Once you have checked the list, tell it which to add and it inserts them in order, in your citation style.

You just gave grand rounds, or a lecture at a conference, or put up a poster. PubMed will not have it, so you hand it to Cowork. Either you drop the line in yourself, or you give Cowork the details and it writes the entry in your CV's format and slots it into the right section. The second way is usually faster, and it keeps a talk you add today matching the ones from five years ago.

Two easy ways to hand over the details:

Give it enough to make a complete entry:

• The title or topic
• What it was: grand rounds, invited lecture, accepted poster, panel
• Where and when: institution or conference, city, date

It shows you the new line before it commits, the same as the papers. You read it, you say yes, it goes in.

Promotion packets and many CVs want short statements on your service, teaching, and research. These are the parts people put off, because starting from a blank page is the hard bit.

You do not start from blank. You give it the raw material and it drafts, then you correct, because your name is on it.

• Tell it what you did, in any order, in plain notes. The committees, the residents you teach, the projects.
• Ask it to draft each section in a few tight paragraphs, in the first person, no inflation.
• Read it as yourself. Cut anything you would not say out loud to a colleague. Fix anything that is not exactly true.

It is faster to fix a draft than to write one. But the judgment stays yours, and so does the final wording.

Here is the real payoff of having it all in one file. You keep one full CV as the source of truth, and Cowork generates whatever shorter version a given form wants, from the same content.

Keep one master CV. Then ask for what you need from it:

• A two-page short CV that keeps only the recent and the relevant
• A short bio in a few sentences, for a talk introduction or a website
• An NIH biosketch in the required layout
• A clean Word file or a PDF, formatted consistently and in reverse chronological order

Each one is a request, not a rebuild. The master file does not change. You get a new version next to it, and you keep the ones you use.

Skip this for a general CV. Use it when a specific thing is on the line, a promotion, a grant, a new position, and the same record should lead with different work.

Tell it what the version is for and what to bring forward. "This is for a teaching-track promotion, so lead with education and put research after." It reorders and trims from the master, and you still read it before it counts. The facts do not change, only which ones come first.

This is the reason to set it up at all. Instead of rebuilding the CV in a panic before a deadline, you let it check for new papers on its own and hand you a short list to approve.

1. Schedule the PubMed check. Tell Cowork to run the Step 2 search once a month or once a quarter, whatever fits how often you publish.
2. Keep it draft-only. Same rule as before. It finds candidates and shows you the list. It does not add anything until you confirm.
3. Approve in a minute. When the list arrives, you check each one is yours and tell it to add the keepers. The CV stays current without a single rebuild.