Headhunting in practice · Sourcing channels
How do headhunters find candidates?
Channels, search and first contact
Headhunters usually start with their own CV database and network and ask people they trust for referrals, then look on recruiting platforms, in professional communities, at industry conferences, and in patents, papers, open-source code, company websites and the news. When they search, they combine every way a keyword might be written, search by company and team rather than by title, which is often more precise, and once they find one good fit, work outward through that person’s colleagues and collaborators. Once the list is ready, they decide who comes first and reach out with a specific reason.
01
When a new search comes in, where should you look first?
When a new search comes in, many people’s first instinct is to open a recruiting platform and type in keywords. Experienced consultants usually pause first and look through what they already have. Past searches, candidates you have spoken to and org charts you have mapped often give you an answer faster than a fresh search, and you have already had some contact with these people.
- Your CV database: go through it by company, by field and by the roles you have put people forward for, not just by job title.
- Org charts you have built: you have already mapped who sits where in the teams at target and peer companies.
- People you spoke to where nothing came of it: someone who did not want to move then may be in a different situation a year or two later.
- Your network: long-standing clients, hiring managers you have worked with and candidates you have placed may all know the right person.
Referrals are the most underrated channel at this stage. Pick a candidate you have placed, someone you trust and who trusts you, explain what the role needs, and ask whether they know anyone working in this field. People who come through a referral often have a closer background, and they start with a little more trust in you. When you ask for a referral, explain how you will approach the person and whether you will mention who referred them, so your referrer feels at ease.
02
Beyond recruiting platforms, where else do headhunters find candidates?
Recruiting platforms are good for finding people who are looking, or open to looking, and they are the most common place to contact candidates. But the people who are truly hard to find rarely update their CVs, or keep them very brief. That calls for a different approach: look where they work, where they speak and where they put their names.
-
Professional communities and industry conferences
Technical communities, industry associations, in-person meetups and conference speaker lists. People willing to get up on stage usually have real depth in their field.
-
Patents and papers
Inventor and author lists show who works on what, which institution they are affiliated with and who they collaborate with. Useful for technical and R&D leadership roles.
-
Open-source code
Contributors to and maintainers of open-source projects, and the authors of technical blogs. You can see what someone has actually written, and how long they have kept at one area.
-
Company websites and news
Team pages, executive bios, press releases, funding coverage and careers pages. They show which team a company is growing and who leads it.
Each of these channels has its limits. A long list of papers does not mean someone has taken a product to mass production or led a team inside a company. Active open-source contributions may be a hobby. A conference title is sometimes written just for the talk. So treat the people you find here as leads first, then check them against other sources such as CVs and team pages to work out who they are, where they are now and what they do.
One more thing to watch: the same person may write their name differently in papers, code and CVs, and different people may share a name. Check carefully before merging records, so you do not treat two people as one, or contact the same person twice.
03
What CV search techniques do headhunters use?
How good your results are depends largely on your keywords. Candidates describe the same thing in many ways on their CVs, so if you only search for the one word in the job description (JD), you will miss a lot of people. You can work through these steps in order:
-
Turn one word into a set
List the keywords in the JD, then add synonyms, abbreviations, Chinese and English versions, and the terms the industry uses. Vision algorithms, for example, might also appear as computer vision (or simply CV), perception, 3D vision or point clouds.
-
Combine them with OR, AND and NOT
Many search boxes accept this Boolean syntax: OR between synonyms in the same set, AND between different requirements, and NOT for anything you want to exclude. For example: (vision OR perception) AND (robotics OR production line) NOT intern.
-
Search by company and team, not just title
The same role goes by very different names from one company to the next: head, manager, director, Tech Lead. Settle on the target companies and teams first, then see who is on those teams. It is often more precise than searching by title.
-
Work outward from one good fit
Once you find someone who fits well, look at their former colleagues, their paper co-authors, their co-inventors on patents, fellow contributors to the same open-source project and other speakers at the same conference.
-
Keep a record of your searches
Jot down which keywords, companies and channels you have searched. When a similar role comes up you can pick up where you left off, and you will know where you have not looked yet.
Here is an example. Say you are working on the Head of Computer Vision role at Xingchuan Robotics (Xingchuan is the humanoid robotics company we use as our running example), and the JD asks for experience with production-line vision and mass production. Searching only for “head of computer vision” gives you very little. List a few peer companies in industrial vision and robotics first, see who on their vision teams has led production-line projects, then work outward through one of those people’s collaborators and former colleagues. Your list will be far more solid.
Company names come in many forms too: the full name, the short name, the English name, subsidiaries and former names. List them all before you search.
04
Once you have a list, who do you contact first and what do you say?
Once the list is ready, do not just call from the top down. First rank it on a few things: fit, whether now is a good time, whether there is a good way to reach them, and whether any restriction puts them off limits.
- Fit: how many of the must-haves they meet, whether they bring any of the nice-to-haves, and which of these are still guesses.
- Timing: how long they have been in their current role, whether their team has changed recently, and what stage the company is at.
- Route: whether a mutual contact can introduce you, or whether you can only reach out directly.
- Restrictions: whether their company is on your off-limits list, and whether you have spoken before and how it went.
For a first contact, three things are enough: who you are, why you are reaching out to them, and what about this opportunity connects to their experience. “I noticed you’ve done a lot of work on production-line vision. I’m working on a role leading the vision team at a robotics company. Would you have ten minutes for a call?” is more likely to get a reply than a generic introduction. Agree with your client beforehand whether you can name them, and how much you can say.
A “no” is still a useful result. You can ask what would make them open to a conversation in the future, or whether they know someone suitable, and then note it down. Silence is an answer too. Wait until you have a better-matched opportunity before getting back in touch, rather than chasing again and again.
05
How can HunterBuddy help you find candidates with AI?
The hard part of these methods is not thinking of them. It is the time they take: going through the CV database, reading the papers, cross-checking one person’s records across several places. It is hard for one person to do all of it for every search. HunterBuddy, the AI partner for headhunters, takes on the searching and checking. Back to the Xingchuan Robotics example: the hiring manager, Qing Xu, has sent over the JD for the Head of Computer Vision role.
-
Checks what you already have
It checks the org charts you have mapped and your CV database first, and finds Yue Lin in an org chart you mapped earlier.
-
Fills the gaps from public sources
Where your database falls short, it looks in papers, patents and open-source communities, and finds Ning Zhou through papers on robotics perception.
-
Checks and merges duplicate records
Yue Lin has a record in your CV database, on a team page and in a paper. Once his identity is confirmed, the three are merged into one person.
-
Spells out the reasons and what is left to confirm
Every candidate comes with the reasons for the match and the sources. It also marks which channels have been searched and which are still untouched.
Xingchuan Robotics · Head of Computer Vision
Candidates worth approaching
Illustrative: the list HunterBuddy hands you. Why each person fits and what is still to confirm are noted alongside.
- Industrial vision · Team leadYue Lin
Has led a production-line vision team. Three records merged. To confirm: team size and reporting line.
- Robotics perception · Senior engineerNing Zhou
Relevant papers and project experience. To confirm: any mass-production experience.
- Computer vision · Project leadSi Chen
Has led cross-team projects. To confirm: current work and openness to a move.
The list is a sourced reference, not a final verdict. Whether someone is open to a move is for you to ask. You decide who to contact first and what to say, and it does not contact anyone on your behalf. CVs downloaded from recruiting platforms can also be uploaded and checked alongside the rest. HunterBuddy is in development, with an invitation-only trial.
06
What are the most common mistakes when finding candidates?
-
Searching only one channel
Not finding someone on a platform does not mean they are not out there. Switch to papers, code, conferences and company websites, and you can often find a different set of people.
-
Searching only by title
Every company names its titles differently, and the same title carries different weight at a large company and a startup. Searching by team and by what people have done is more precise.
-
Mistaking good on paper for a good fit
Many papers, active open-source work and an impressive title are all just leads. How big a team someone has led, and whether they have done mass production, you only find out by talking.
-
Contacting the same person twice
When your CV database, a recruiting platform and a paper each hold a record that never gets merged, you end up reaching out twice, and it costs you in their eyes.
And one bottom line: find people using only public professional information and what candidates give you themselves. If someone does not want to talk, or does not reply, that is their choice. Respect it, and you keep the chance to reach out again later.
FAQ
What people ask
about this.
What channels do headhunters use to find candidates?
The common ones are your own CV database and the org charts you have built, your network and referrals, recruiting platforms, professional communities and industry conferences, patents and papers, open-source code, and company websites, team pages and news. Usually you start with what you already have, then fill the gaps from outside. Treat people found through papers or code as leads, and contact them only after checking against other sources.
How can headhunters search CVs more effectively?
Expand each keyword in the JD into a set of synonyms, abbreviations and Chinese and English variants, then combine them with OR, AND and NOT. Searching by target company and team is often more precise than searching by title. Once you find a good fit, work outward through their former colleagues and collaborators, and keep track of which keywords and channels you have already searched.
How do you find candidates through patents, papers and open-source code?
Settle on keywords for the field, then look up the inventors on relevant patents, the authors of papers and the contributors to open-source projects. Look at the organizations they are affiliated with, their collaborators and how long they have worked in the area. This only shows that someone has built up experience in a field. It does not show they have done mass production or led a team. Check anyone you find against their CV, the company website and other sources to confirm it is the same person and where they are now, before deciding whether to contact them.
How can you use AI to find candidates?
AI is well suited to searching and checking: going through your CV database, reading papers and patents, matching up one person’s records from different places, and listing the reasons for a match with their sources. An AI assistant for headhunters such as HunterBuddy checks your existing org charts and CV database first, then fills the gaps from public information, and marks anything uncertain “To confirm”. Who to contact first, and what to say, is still your call.