Boolean Search for Founders: Finding Co-Founders and Advisors Without a Recruiter

Recruiters spend a surprising proportion of their working life writing search strings. Founders running their own co-founder or advisor search usually do not, and it is the largest single source of wasted time I see: hours spent scrolling a results page that a fifteen-minute string would have reduced to forty names.

This article covers the mechanics of Boolean search as it applies to founder-level searches, then, more importantly, the point at which Boolean stops being the right tool.

The operators, briefly

Three practical notes. Operators must be capitalised on most platforms. Parentheses matter more than people expect, because without them the engine applies its own precedence and quietly returns something other than what you asked for. And NOT is the operator that most often destroys a search: exclude too eagerly and you remove the strongest candidates along with the noise.

The four-block method

Rather than writing one long string and debugging it as a whole, build four blocks and test each independently. This makes it obvious which part of the query is destroying your result count.

Block 1: role synonyms

The most common failure in founder-run searches is searching for the title you have in mind rather than the titles the market uses. The person you want to be your technical co-founder very often does not have CTO anywhere on their profile.

("CTO" OR "Chief Technology Officer" OR "VP Engineering" OR "Head of Engineering" OR "Engineering Manager" OR "Principal Engineer" OR "Staff Engineer" OR "Tech Lead" OR "Lead Engineer" OR "Founding Engineer")

Note the inclusion of staff and principal engineer. In many strong engineering organisations those are the people with the deepest technical judgement, and they are systematically missed by founders searching for leadership titles.

Block 2: evidence

This block encodes the Deciding criterion from your brief. It should describe something the person did, not something they claim to be.

("SOC 2" OR "ISO 27001" OR "PCI DSS" OR "PSD2" OR "DORA" OR "penetration test" OR "security review" OR "regulatory reporting")

Block 3: context

(fintech OR "financial services" OR banking OR payments OR lending) AND (startup OR "Series A" OR "Series B" OR "early-stage" OR B2B OR SaaS)

Block 4: exclusions

NOT (recruiter OR recruitment OR "talent acquisition" OR headhunter OR intern OR student OR "looking for opportunities" OR coach OR mentor OR "business development representative")

Combine them only after each block has been tested on its own. A useful diagnostic: if adding the evidence block drops your result count by more than 95%, your evidence terms are too narrow or too jargon-specific, and you should widen them before assuming the market is empty.

X-ray search: reaching profiles through a search engine

Platform search results are shaped by the platform's commercial interests and by your subscription tier. X-ray search goes around that by querying a general search engine for public profile pages directly.

site:linkedin.com/in ("CTO" OR "Head of Engineering") AND fintech AND ("SOC 2" OR "PSD2") AND (Berlin OR Munich OR Hamburg)

Variants of the same pattern work on other public sources, and they are frequently higher signal than a profile database for technical searches:

• Code hosts. Search repository topics, contributor lists and issue discussions in your problem space. The person who has argued thoughtfully in an issue thread about the exact constraint you face is a better lead than any title match.
• Conference programmes. site:conference-domain.com speakers 2024 OR 2025 plus your topic. Anyone who has given a talk has publicly demonstrated both the experience and a willingness to explain it.
• Engineering and industry blogs. Search for the named authors of posts about your constraint. They are usually still reachable and often flattered to be asked.
• Academic and patent databases, for deep-technology searches where the relevant expertise never appears on a professional profile at all.
• Professional registers and regulator lists, for advisor searches in regulated industries.

Where Boolean search breaks down

Now the important part. Boolean search matches strings. It has no notion of meaning, so it fails on exactly the criteria that decide a co-founder or advisor search.

The standard response to this is to keep widening the string, then manually read four hundred profiles. That works, and it is how the search profession spent twenty years. It costs roughly six to ten hours per shortlist of fifty, which is a serious tax on a founder who is also building a company.

The alternative is to split the query in two. Use hard filters for the facts that are genuinely facts, and a plain-English description for the experience that has to be inferred:

• Hard filters: seniority, geography, industry, company size, current and past companies, years of experience, exclusions.
• Contextual description: has been the most senior engineer in a team of five or more at a company selling software into banks, and has personally handled a customer security review.

The second half is not expressible as a Boolean string, but it is expressible in a sentence, and it is what actually separates a shortlist of fifty from a results page of four thousand.

Six ready strings for founder searches

Adapt the geography and industry terms. Each of these is a starting point to be narrowed, not a finished query.

("CTO" OR "VP Engineering" OR "Head of Engineering" OR "Principal Engineer" OR "Staff Engineer") AND (B2B OR SaaS OR "enterprise software") AND (Stockholm OR Copenhagen OR Helsinki OR Oslo) NOT (recruiter OR "talent acquisition" OR consultant)

("VP Sales" OR "Head of Sales" OR "Chief Revenue Officer" OR "Sales Director") AND ("product-led" OR "self-serve" OR PLG) AND (enterprise OR "mid-market") NOT (SDR OR "business development representative" OR recruiter)

("regulatory affairs" OR "quality assurance" OR "RA/QA" OR "notified body") AND ("CE mark" OR "MDR" OR "510(k)" OR "FDA clearance") AND ("medical device" OR "digital health" OR SaMD)

("Founding Engineer" OR "Staff Engineer" OR "Principal Engineer" OR "Machine Learning Engineer") AND (LLM OR "large language model" OR RAG OR "inference" OR "evaluation") AND (production OR "in production" OR shipped) NOT (student OR intern OR "PhD candidate")

("procurement" OR "supply chain" OR "purchasing" OR "commissioning") AND (hospital OR NHS OR "health system" OR "trust") AND (director OR head OR lead) NOT (vendor OR supplier OR sales)

("pricing" OR "monetisation" OR "monetization" OR "packaging") AND (SaaS OR API OR "usage-based" OR "consumption") AND (head OR director OR VP OR lead) NOT (recruiter OR "pricing analyst" OR retail)

How to keep the output usable

A shortlist that lives in a browser tab is not a shortlist. Whatever tool produces the names, the output needs a structure you can work from.

1. 1One row per person, with the profile link, the evidence link, and the specific Deciding criterion they satisfy.
2. 2A confidence column. High means you have direct evidence, medium means an inference, low means it looks plausible. Contact the high rows first.
3. 3A wave column, so you can send in batches of fifteen and improve the message between waves.
4. 4A status column with date of contact and reply. Without dates you will lose track of who to follow up.
5. 5A rejected column with a reason. In two months you will re-run this search and you need to know why you passed on someone.

Once the list exists, the constraint moves from finding people to writing messages they will answer. That is covered in Cold outreach to co-founders and advisors. If you have not yet written the brief the string is meant to encode, start with How to find a co-founder.