# Hiring Assistant 2 Crowds the Same Engineers. Source the 41% in German

_LinkedIn Hiring Assistant 2 rewards verified, connected-app profiles. Here is who it under-surfaces, and how to source engineers outside LinkedIn's view._

- Canonical URL: https://www.refolk.ai/blog/hiring-assistant-2-connected-apps-source-engineers-outside-linkedin
- Published: 2026-11-18
- Reading time: 9 min
- Keywords: LinkedIn Hiring Assistant 2, LinkedIn connected apps GitHub recruiters, Hiring Assistant 2 limitations, source engineers outside LinkedIn, LinkedIn Recruiter AI agent November 2026

LinkedIn Hiring Assistant 2 reaches English-language customers this month, and it comes with a number every recruiter will repeat: members with a connected app get up to 2.9x more recruiter messages. Read that as a warning, not a tip. If the agent keeps pointing thousands of recruiters at the same verified, connected, responsive engineers, those inboxes fill up first. The engineers it can't see well are where your outreach still gets read.

## What changes in LinkedIn Hiring Assistant 2 this month

Hiring Assistant 2 adds recruiter-level memory and puts more weight on verified identity, Connected Apps and responsiveness when it ranks candidates. LinkedIn announced it on October 1, 2026, presented it at Talent Connect 2026 in New York, and says it is free for English-language customers from November 2026.

One caveat on rollout. Some sources describe it as free for *existing* English-language Hiring Assistant customers, so don't assume every Recruiter seat gets it automatically. Check your contract before you plan around it.

Here is what is actually new for sourcing:

- **Recruiter memory.** The agent learns from what you hired, who you invited and who you passed on, then shapes the next search to match.
- **Receptiveness signals.** It scores candidates on platform engagement, Open to Work details and InMail responsiveness.
- **Verified identity.** LinkedIn says 115 million members have verified their identity to date.
- **Connected Apps.** An opt-in menu that currently lists 26 partners, including GitHub, Replit and Behance, with Chess.com, ElevenLabs and Medium announced as new partners at the keynote. It pulls verified candidate data from those sources into Hiring Assistant's evaluations.
- **ATS integration.** At launch it connects to Greenhouse, Lever, SmartRecruiters, Zoho Recruit and Tracker. iCIMS and Workday are listed as coming.

The scale matters. LinkedIn says more than 20,000 companies now use its AI hiring agents, and that recruiters on v1 reviewed 83% fewer profiles before finding a qualified match. Reviewing fewer profiles is the point of the product. It also means the agent decides more of what you see.

## Why the 2.9x figure is a crowding signal

The 2.9x number tells you where recruiter attention already piles up, not who the best engineers are. It is an "up to" figure from LinkedIn, and it isn't controlled for anything.

> **2.9x** More recruiter messages for LinkedIn members with a connected app
>
> LinkedIn's own "up to" figure, uncontrolled for how active those members already are.

Think about who opts into Connected Apps. Connecting a GitHub account to a LinkedIn profile is something an engaged, profile-tending, often job-curious member does. Those are the same traits the receptiveness score already rewards: platform engagement, Open to Work and InMail replies. So the 2.9x is mostly correlation. The people with connected apps were probably getting more messages anyway, and the new weighting pushes more on top.

For a recruiter, the practical read is simple. A candidate the agent ranks highly for you is likely ranked highly for other recruiters too. Your InMail lands in a queue. The extra messages don't create more strong engineers. They split the same ones among more recruiters.

## How recruiter memory narrows the shortlist

Recruiter memory is the agent's record of your past hiring choices, used to rank future candidates. It makes each search look more like your last one, and across many companies it makes shortlists overlap.

LinkedIn's own example is a product manager search. The agent will look for someone close to the office, in the retail industry and at mid-level, "because those are all the things you've hired for in the past." One summary of the release puts it more bluntly: "Who you invited before and who you passed on shapes what it suggests next."

That is a feedback loop, and it has two effects LinkedIn's materials don't spell out:

1. **Your own pool shrinks.** Each accepted suggestion strengthens the pattern, so profiles that don't fit your history (a different industry, an unusual title, a non-English profile) rank lower over time.
2. **Your competitors' pools converge with yours.** Recruiters at similar companies with similar hiring histories get trained toward the same profiles. Two fintechs hiring backend engineers in the same city will see overlapping shortlists, and they will message the same people in the same week.

At the keynote, LinkedIn said customers wanted Hiring Assistant to "just be smarter." Smarter about *your* history is useful for screening volume. It is a liability for hard-to-fill engineering roles, where the person you need often looks nothing like your last three hires.

The cost shows up as time. When every recruiter's agent hands back the same verified few, replies dry up, searches stall, and a senior role drifts while you keep re-messaging a contested shortlist.

## Hiring Assistant 2 limitations: who the agent can't see well

The agent works best on engineers who are on LinkedIn, verified, active, connected and titled in English. Everyone outside that profile is under-ranked or invisible, and that is a large group.

Here are the main blind spots:

- **Engineers with no LinkedIn profile.** Connected Apps can't add an engineer, researcher or patent holder who has no LinkedIn profile at all. (That point comes from Pin, a competing vendor, but it follows directly from how the feature works.)
- **Engineers who never connected GitHub.** Connected Apps is opt-in, and GitHub joined the partner list after the June launch. An engineer who hasn't opted in contributes nothing to that signal.
- **Senior engineers whose work is private.** Someone who spent years in a company's private repos has nothing public to connect. As one analysis warns, a missing credential "should not silently become evidence of incompetence." A ranking model weighted toward verification can still treat it that way.
- **Engineers with non-English titles.** They are on LinkedIn, just not where an English-tuned search looks. More on this below.
- **Engineers who never applied to you.** The ATS integration brings in past applicants from Greenhouse or Lever. That widens the pool to people you've already met, not people you haven't.

Compare the size of the pools. GitHub's Octoverse 2025 report says more than 180 million developers build on GitHub, and more than 36 million joined in a single year. Connected Apps only sees the ones who linked that account to a LinkedIn profile. The open web is the bigger pool, and the agent doesn't look there.

## The German test: 41% of the pool uses German titles

In Refolk's index of professional profiles, 41% of Germany's software engineer pool uses German-language titles that a search for "Software Engineer" won't match. Language is a blind spot on the scale of the GitHub one, and these engineers are already on LinkedIn.

> **41%** Share of Germany's engineer pool using German-language titles
>
> From Refolk's index: 27,105 of 66,023 profiles say Softwareentwickler, Entwickler and similar.

The numbers, from Refolk's index:

| Segment | Profiles |
|---|---|
| US: Software / Senior / Staff Software Engineer | 562,602 |
| Germany: same English titles | 38,918 |
| Germany: German-language titles (Softwareentwickler, Software-Entwickler, Softwareingenieur, Entwickler) | 27,105 |
| Germany: combined pool | 66,023 |

Two things stand out. First, adding the German titles grows the German pool by 70% over the English-title count. Second, the US English-title pool is 14.5x the German one, which is why a model tuned on English-speaking recruiters' histories will tend to treat English titles as the norm.

One honest caveat: the bare title "Entwickler" pulls in some non-software and founder profiles, so 27,105 is an upper bound. Even trimmed, it's a large share of the market.

### Who sits behind each title

The two groups work at different kinds of companies. In small samples from Refolk's index (25 profiles each, so treat this as directional):

- **German-title engineers** cluster at Mittelstand and regional employers like mediendesign AG, TOPAS Industriemathematik, envelio, SAP and Giesecke+Devrient, around Nuremberg, Bremen and Dresden.
- **English-title engineers** cluster at Berlin startups like Langfuse, Helsing, dltHub and Hex-Rays.

So the English-title, Berlin-startup engineer is the one Hiring Assistant 2 surfaces easily, and the one every Berlin recruiter is already messaging. The Softwareentwickler at a Dresden or Bremen industrial software company is on LinkedIn and less contested. Getting to them takes a search that thinks in both languages. That's the gap [Refolk](https://www.refolk.ai/) closes: you describe the engineer you want in plain English, and it matches across title variants, languages and sources instead of one exact string.

> **Example search:** "Senior backend engineers in Germany outside Berlin, including profiles titled Softwareentwickler or Softwareingenieur, with public Go or Rust work on GitHub" - [run it on Refolk](https://www.refolk.ai/s/2bykqcxx2v)
>
> A ranked shortlist that merges English- and German-titled profiles with their GitHub activity, so the Mittelstand engineers show up next to the startup ones.

The same logic likely applies in any market where engineers write their titles in the local language: some slice of the pool sits where an English-tuned agent under-ranks it. Germany is just the case the index measures cleanly.

For a sense of how contested the English-titled Berlin pool is, this is the open demand there right now:

> **Engineering roles open in Berlin** - [see the open roles](https://www.refolk.ai/jobs/search?family=engineering&location=berlin)

## How to source engineers outside LinkedIn's agent

Let Hiring Assistant 2 handle the verified, active, English-titled pool, and spend your own sourcing time on the people it ranks low. Those are the inboxes with room in them.

A practical playbook:

1. **Split the work.** Use the agent for volume roles and for candidates who are clearly on the market. Don't use it as your only source for senior or niche engineering roles.
2. **Search every title variant.** In Germany that means Softwareentwickler, Software-Entwickler, Softwareingenieur and Entwickler alongside the English titles. Do the same for every non-English market you hire in.
3. **Go to GitHub directly.** Look at commit history, maintainers and frequent reviewers on the repos your team depends on. An engineer with no Connected App may still have years of public work.
4. **Read the open web.** Conference talks, engineering blog posts, package maintainers and paper authors often have thin or missing LinkedIn profiles.
5. **Don't penalize a missing badge.** No verification and no connected GitHub usually means private work or low LinkedIn activity, not weak skills.
6. **Write outreach that shows you looked.** Engineers outside the crowded pool get fewer messages, so a specific note about their actual work stands out more.

The bigger point is about competition. When more than 20,000 companies use the same agent tuned on similar signals, the advantage goes to the recruiter whose sourcing looks *different*. Message the rest.

## FAQ

### Is LinkedIn Hiring Assistant 2 free and automatic for all Recruiter customers?

LinkedIn says Hiring Assistant 2 is free for English-language customers from November 2026, and it was announced on October 1, 2026. Several sources word that as free for *existing* English-language Hiring Assistant customers, and I couldn't confirm that every Recruiter seat gets it automatically. Check with your LinkedIn account team before assuming you have it.

### Should engineers connect GitHub to LinkedIn to get more recruiter messages?

LinkedIn says members with a connected app get up to 2.9x more recruiter messages, so connecting GitHub will likely increase volume. For recruiters, though, that same number means connected engineers are the most contested candidates on the platform. The figure isn't controlled for how active those members already were, so read it as a sign of crowding, not a quality signal.

### Can Hiring Assistant 2 find engineers who aren't on LinkedIn?

Only if they have already applied to you. The ATS integration with Greenhouse, Lever, SmartRecruiters, Zoho Recruit and Tracker brings in past applicants, but Connected Apps can't add anyone without a LinkedIn profile. With more than 180 million developers on GitHub, many engineers sit outside the agent's view entirely, so to source engineers outside LinkedIn you need GitHub and open-web search.

---

_From the Refolk blog. More posts: https://www.refolk.ai/blog/index.md_
