> ## Documentation Index
> Fetch the complete documentation index at: https://guide.kello.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# How Kello evaluates candidates

> Kello reads each profile the way a senior recruiter would: it starts with your company, reads careers for meaning, and weighs whether people would say yes.

Kello reads each candidate's profile the way a senior recruiter would, for meaning rather than keywords. It starts with your company, reads what each person's career actually says, and weighs whether your role is one they would say yes to.

Kello doesn't run a filter over job titles. It reads.

## Where does Kello start?

With your company. Before it looks at a single candidate, Kello reads what your company does, its history and growth, its current team, and the kind of people it has hired. It learns that last part from the public professional profiles of people who joined you.

That gives Kello a picture of who tends to succeed with you, before your JD (Job description: the document that describes the role.) says a word.

## How does Kello read a profile?

The way a senior recruiter would: unstructured, for meaning. For each person, Kello looks at:

* What each of their employers actually does, not just the name.
* How big each company was, and what stage it was at, when the person joined and when they left.
* Whether they were promoted early compared with their peers.
* What their team looks like.
* What they worked on.

For mid-level managers, Kello also looks at how their team performed. For senior roles, it looks at how their org or division performed.

## What does that look like on one profile?

> [!NOTE]
>
> This is an illustrative, fictional work history. It is not a real person.

Say you're hiring a Head of Engineering for a Series B startup. One profile reads:

| Company | Role | Years |
| - | - | - |
| A payments company | Software Engineer | 4 |
| A B2B software company | Senior Software Engineer, then Engineering Manager | 4 |
| A large consumer internet company | Engineering Manager | 2 |

> [card] A keyword search
>
> Sees "Engineering Manager" and stops there.

> [card] Kello
>
> Reads further.
>
> * It checks what the payments company does and finds it sells to other businesses, so this person built for business customers from the start.
> * It sees that the payments company had a few dozen people when they joined and several hundred when they left, so they lived through a growth phase.
> * At the B2B software company, they moved from Senior Engineer to Engineering Manager faster than most of their peers there.
> * Kello looks at the team they managed and how it performed.

Then it asks whether this person would say yes. Moving from manager at a large company to Head of Engineering at a growing startup is a step up in title and scope. That makes them worth showing you. If your role were Senior Engineer, Kello would leave them out, because it would be a step down.

On your list, that reading becomes plain words. **Why this candidate** might say they've built for business customers and grown with a company through a scale-up. **Worth checking (The one thing on each candidate Kello suggests you confirm.)** might suggest asking whether they still write code day to day.

## Would they say yes?

A strong profile is only useful if the person would consider your role. Kello leaves out people for whom your role is a step down in title, scope or company. It also weighs how likely each person is to move.

Nobody on your list applied, and most aren't actively looking. Nobody is contacted until you mark them Yes or Maybe.

## Why does the list have different kinds of people?

Most roles can be done well by people with different backgrounds. Kello looks for several different kinds of people who could succeed, so your list isn't one profile repeated 20 times. Your Yes and No marks then show Kello which kinds your team prefers.

## Does Kello look at public work?

For engineering and AI roles, yes, when the role calls for it. Kello looks at code shipped on GitHub, research published, models built (for example, on Hugging Face), hackathons and product launches.

> [!NOTE]
>
> Missing accounts are never held against anyone. Plenty of strong engineers have no public code.

## Does Kello read some roles differently?

Yes. Kello has function-specific reading for AI-native engineers, forward-deployed engineers (FDE) and engineering roles. More functions are coming.

Kello takes roles in every function, entry level to C-suite.

## What does Kello never consider?

> [!TIP]
>
> Kello never considers age, gender, religion, race, marital status or similar personal traits. If a brief asks for one, Kello leaves it out rather than swapping in a stand-in, such as an experience cap for an age.

Kello doesn't search on personal traits. It looks across many kinds of companies, schools and career paths. See [fairness](https://guide.kello.ai/kello/trust.md#what-does-kello-never-consider).

## Where does the information come from?

Every insight on your list traces back to the candidate's own profile text or Kello's data on companies and people. Kello never invents them. Candidate data comes from public professional information. See the [privacy policy](https://kello.ai/privacy-policy).

## Related

> [card] [Your list](https://guide.kello.ai/kello/your-list.md)
>
> Where this reading shows up.

> [card] [Data, privacy and fairness](https://guide.kello.ai/kello/trust.md)
>
> Professional signals, never personal traits.

> [card] [Write a great brief](https://guide.kello.ai/kello/write-a-great-brief.md)
>
> Show Kello who you'd hire.

Send your next role to [source@kello.ai](mailto:source@kello.ai).


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.