Write the profile for how recruiters actually search

How recruiters actually find people

The workflow is search, scan, open, skim. They search on concrete terms — a language, a framework, a title, a location, sometimes a school. The results list shows your name, headline, and location, and nothing else. That headline is doing all the work of getting you opened.

So a headline of "Aspiring Software Engineer | Passionate about technology" fails twice: it contains no term anyone searches for, and it says nothing that would make someone click.

Two structural facts change how you should write. First, the search is largely boolean over text: a recruiter types something like python AND react AND intern with a location and a school filter, and you either contain those strings or you do not. Second, recruiter outreach is rationed — the messages they can send are a limited monthly resource, so every one they spend is a judgement about likely reply. A profile that looks abandoned costs them a message with no return, and they know it.

Nothing on your profile is read in isolation. It is read as the twelfth of forty profiles, at speed, by someone deciding where to spend a message. Every section below should be judged by that comparison, not by whether it reads well on its own.

The headline

[what you are] · [proof or specialisation] | [concrete stack] | [what you are building or seeking]

Concrete beats aspirational. Load it with the nouns a recruiter would type — the language, the framework, the domain, the credential — while still reading like a human wrote it. You have around 220 characters; use most of them. The first roughly 60 characters are what shows in a truncated mobile results list, so the searchable nouns go early.

• Include your actual stack in the words the industry uses. Write JavaScript, not JS; write PostgreSQL, not Postgres, if the postings you want say PostgreSQL.
• Include one credential or proof point if you have one: a publication, a shipped product, a named employer, a graduation year.
• Include what you are looking for if you are looking. Recruiters filter on open-to language, and a headline that names the role you want is doing filtering work for them.
• Include location or relocation willingness if you are geographically flexible. It is a hard filter on their side and a cheap sentence on yours.
• Avoid: passionate, aspiring, enthusiast, ninja, guru, rockstar, hustler. They occupy space and signal nothing.
• Avoid emoji dividers as separators between every clause. One or two are fine; a row of them eats characters you need for nouns.

The About section, worked end to end

LinkedIn truncates About behind a see-more link at roughly 250 to 300 characters on mobile. Whatever is above that fold is the entire section for most readers, so the first sentence must be the summary of you, written in plain declarative form. No throat-clearing, no story opener, no quote.

The structure that works, in order:

1. One declarative sentence. What you are, where, and the single most credible fact about you.
2. One sentence of proof with a number in it. Above the fold if you can manage it.
3. What you actually do, two or three short lines. Concrete work, not adjectives.
4. The stack, as a scannable line. This is also index food.
5. One or two specifics with links — the project, the paper, the product.
6. What you want next, stated plainly, including constraints that matter.
7. How to reach you. An email address in text. Many recruiters will not spend a message on someone who has given them a cheaper channel.

WORKED EXAMPLE (illustrative) I am a CS undergraduate at San Jose State building applied AI and automation tools, and a co-author on a peer-reviewed SIGCSE 2026 work. Most recently I built an NLP scoring pipeline over 50+ market sources per day that cut manual research time 60%. What I actually do: take a messy, deadline-driven workflow, turn the unstructured part into records, and put an automated quality gate in front of anything that leaves the building. Stack: Python, Java, JavaScript, TypeScript, React, SQL, RAG, local LLMs (Ollama), Git, pandas, scikit-learn. Things you can click: - FreightDesk, an AI back office for small freight brokerages - DIRA, a zero-dependency security scanner (MIT) - 60 technical guides, all free Looking for: Summer 2027 SWE or applied-ML internships, Bay Area or remote. Reach me: name@example.com

Notice what is absent: no paragraph about passion, no third-person narration, no wall of text. The mobile app renders long paragraphs as a grey slab that nobody finishes, and the sections that get read are the ones with white space around them.

Experience entries are not job descriptions

Copying your duties from the job posting is the most common waste of space on LinkedIn. Write what changed because you were there. Action, object, method, result — with the result quantified when it honestly can be. The first entry is real work; the rest are illustrative examples of the shape.

1. REAL Before: Responsible for maintaining internal reporting tools After: Built an NLP scoring pipeline over 50+ market sources per day, cutting manual research time 60% 2. ILLUSTRATIVE Before: Worked closely with cross-functional stakeholders After: Ran the weekly triage between support and engineering; cut median bug-to-fix time from 11 days to 4 3. ILLUSTRATIVE Before: Assisted with social media After: Wrote and scheduled 3 posts a week for 8 months; grew the club account from 210 to 1,400 followers 4. ILLUSTRATIVE Before: Tutored students in computer science After: Tutored 40+ students through CS1 and CS2; rebuilt the recursion module after 3 semesters of the same failure pattern 5. ILLUSTRATIVE Before: Treasurer After: Managed a $22k annual budget across 10+ events; moved the books off a shared spreadsheet onto a reconciled monthly close

Two mechanical points people miss. LinkedIn lets you attach media and links to an individual Experience entry, which is the only way a specific role gets a clickable artefact next to it. And it lets you associate skills with a specific role, which is what feeds the skill-to-experience matching in recruiter tooling. Both are free and both are usually empty.

Skills, and how the index actually uses them

• The Skills section is search index food. A term listed there is findable even when it does not appear in your prose. That is its primary job, not the endorsement count.
• You get up to 50. Use them, but use them on real terms: languages, frameworks, databases, tools, methods, and domains you could be asked about.
• Only the top three are pinned and visible without expanding. Choose those three deliberately — they should be the three you most want to be hired for, not the three with the most endorsements.
• Attach skills to individual roles and projects. A skill tied to an Experience entry carries more weight in matching than a free-floating one, and it also gives the reader a reason to believe it.
• Use canonical spellings. LinkedIn's skill entries are a controlled vocabulary; pick the suggested entry rather than typing a free-text variant, or you end up indexed under something nobody selects in a filter.
• Endorsements are weak signal. They affect ordering slightly and persuade almost nobody. Do not run an endorsement-trading campaign; it is the lowest-return activity on the platform.
• Prune the embarrassing ones. Microsoft Word on an engineer's profile is not neutral, it is a signal about your judgement of what is worth listing.

Featured, as a portfolio surface

The Featured section is the only place on the profile where a link gets a real image card with visual weight, sitting directly under About where the eye is already going. It is the highest-value real estate on the page and it is empty on most profiles.

Give each featured item a real title and a one-line description. The auto-generated preview title is often a truncated page title, which reads as neglect. What belongs behind those links — and what makes a project worth clicking at all — is covered in the portfolio guide.

The settings checklist

That is roughly half an hour, once, for the entire mechanical layer. It is the highest ratio of outcome to effort anywhere in a job search, and it is the part most people skip in favour of rewriting the About section for the fourth time.

A posting cadence that is actually sustainable

You do not need to post daily. But a profile with zero activity looks abandoned, and a profile with a handful of substantive posts about what you built reads as evidence. The cadence below is designed to be survivable during a term with coursework, which is the only kind of cadence that gets followed.

What to post, in descending order of value per minute: the thing you just shipped, with a link and one paragraph on what broke. A short teardown of something you learned the hard way. A resource you made that other people can use. A milestone with the number attached. The thing that does not work is generic advice with no artefact behind it — everybody already has enough of that in their feed.

Almost all of it is the same content as your portfolio and your guides, so it costs you nothing new to produce. If you write technical content anyway, the discoverability side of that is covered in the AI search guide.

Connection and outreach etiquette

The rules are boring and the violations are expensive. Do not pitch in the first message. Personalise the first line with something specific enough that it could not have been sent to anyone else. Ask a question that can be answered yes or no, because open-ended asks require effort and effort is what does not happen.

CONNECTION NOTE (alum, no prior contact) Hi [name] - fellow SJSU CS. I saw you moved into platform work at [company]. I am building in a similar space (link in my featured section) and would like to follow what you post. No ask. REPLY TO A RECRUITER, EVEN A BAD FIT Thanks for reaching out. That role is not a fit for me right now because [one clause]. If something opens on the [specific] side, I would be glad to hear about it. Here is my resume for your files. ASK, AFTER A REAL EXCHANGE HAS HAPPENED Would you be open to a 15-minute call about how your team thinks about [specific thing]? Happy to work around your calendar, and happy to just ask over message instead if that is easier.

• Reply to every recruiter, including the irrelevant ones. Recruiters move companies constantly and they remember who answered.
• Do not send a connection request and a pitch in the same breath. It is the single most-ignored pattern on the platform.
• Do not ask for a job in a first message to a stranger. Ask a question you actually want the answer to.
• Do not send voice notes to people you have never spoken to.
• Follow up once, after about a week, and then stop. Two follow-ups is diligence; four is a reason to be blocked.

The metrics worth watching

Check these monthly, not daily. The numbers are noisy week to week and the temptation to react to noise is the same failure mode as reading a scale every morning — a problem worked through properly in the signal-versus-noise guide, which is about weight but is really about trends.

The profile is one half of a pair. The document it hands off to is covered in the ATS resume guide, and the search this all feeds — especially if you need work authorisation — is laid out in the internship guide.

If you would rather have it done for you, I rewrite profiles end to end in 48 hours: the LinkedIn Profile Rewrite Sprint →

Tools referenced in this guide

• LinkedIn Profile Rewrite Sprint — a complete rewrite delivered in 48 hours.
• Resume rebuild — the ATS-safe document that pairs with the profile.
• ATS resume guide — what breaks a resume parse and what does not.