Skip to content

The first world model
for people.

Language models read text. Our people world model understands how companies actually work: who builds what, who succeeds where, who is about to leave and who you need next. Orbit is the first, and it is built for hiring.

Hiring manager, PlatformOrbit

Who should we hire next for Platform?

Your team has four people who start projects and nobody who finishes them. Here are three who do, and none of them work at the brands everyone else is searching.

From Ashby, HiBob and Linear
People teamOrbit

Why is our Data team losing people?

Two leads left after the reorg took their scope. Pay is in line. Expect the pattern to repeat.

CEOOrbit

Can we hit the Q4 launch with the team we have?

Not as it stands. Platform is two senior backend short, and those hires take you 70 days. 58 remain.

Built by the people who built hiring engines atMetaSnapAirwallexMcKinsey & Company

Recruiting AI made search faster.
It didn't make hiring better.

Not the problems in vendor decks. The ones talent leaders raise once the demo is over.

01

Everyone searches the same data.

Every tool sells the same people to every company in your market, so you all chase the same names.

02

None of them know your company.

They match a CV to a job description. Your team, your leaders and what has worked for you never come into it.

03

Your hiring trains your competitors.

What you teach a shared tool about good candidates improves the profiles your rivals search next.

04

The best people are screened out unseen.

Thin profiles get ranked down, so the strongest candidate often never reaches a human. Only 21% of recruiters and hiring managers are very confident their systems are not rejecting qualified people.

Greenhouse, 2025 AI in Hiring Report

05

Scores nobody can defend.

A number you cannot explain is a number a hiring manager will not trust, and it is gone the moment they ask why.

06

Faster, and no more certain.

72% of talent professionals say AI made discovery faster. A third say their confidence in finding the right person has not moved.

TestGorilla, Recruiting in 2026

A language model reads text.
A people world model understands companies.

It is trained on how companies really work: how teams are built, how careers move, how work gets done and who succeeds where. Then it learns yours, privately.

Same question
A general AI
A people world model
Who should we hire next for Platform?
I don't have data on your Platform team. In general, look for relevant experience and a mix of senior and mid-level engineers.
A senior engineer who finishes things. Your team has four people who start projects and nobody who lands them. Here are three who do.
Is this candidate right for us?
From the CV, they match most of the job description. I can't tell how they would fit without knowing your team.
Strong from zero to one, weaker at scale. At your size, people with this path thrive for about 18 months. Ask how they handled their last team doubling.
Where should we hire our next ML team from?
Top AI labs and big tech companies are common sources of ML talent.
Two teams at a chip start-up just lost their leads. Eleven people there match your team's DNA, and nobody else is looking yet.

What becomes possible.

Hire for the team, not the job ad

Shortlists judged against who is already on the team and what it is missing.

Find the talent the market has mispriced

Proven people, outside the brands every other tool is bidding up.

Map any team, anywhere

The people who built a product, who stayed, who left and where they went.

Know why people leave

Read from what changed around them, not from what they said on the way out.

Staff the roadmap before it slips

Every goal matched to the team that must deliver it, against real time to hire.

Ask from anywhere you work

The same answers inside Claude, ChatGPT or Glean, signed in as you.

How our models work →

Meet Orbit.
First, it fixes hiring.

Our first model is a recruiter's right hand, not a replacement. It does the reading, searching and chasing, so your team can do the judgement only people can.

Explore Orbit
Staff Platform EngineerSlate ready
Role appears in Ashby
0s
Screen-ready slate
60s
6 screen-ready, each with the reason

From job post to slate in 60 seconds

Orbit starts the moment a role goes live, and nothing is sent until a person says so.

JW
James Walker
Tech lead, inference · Lumen Freight
Undervalued

Five independent signals of the work, none of them a brand. Fills Platform's scaling gap.

ScalerRising fastFills a gap

The candidate the market mispriced

Strong evidence, low visibility. The hire everyone else's tools rank down.

Team DNAPlatform · 14 people
Starts thingsOn target
Scales thingsNeeds more
Goes deepOn target

The shape of every team

Who starts, who scales, who multiplies, and which team is about to stall.

Private by design.
A strategic advantage by default.

Your model compounds with every hire. What it learns about your company stays yours, and never helps a competitor hire against you.

Read our security approach

One model per company

Trained on your data alone. No shared pool.

Hosted in the UK, EU and US

Your data stays in region.

Evidence behind every signal

No black boxes. Every read shows its working.

A person always decides

Nothing is rejected or sent automatically.

Joshua Smith, founder of Halloway
Joshua Smith
Founder, Halloway
Previously Meta, Snap,
Airwallex and McKinsey

Built by someone who's led frontier hiring at Meta and Snap.

Recruiting is moving past keyword search toward models that read teams, not resumes. The next decade belongs to systems that understand how your company actually works and who succeeds inside it.

Halloway is building that layer. Orbit is our first people world model: private to you, built for the hiring calls that shape what your company becomes.

Orbit is a force multiplier, not a replacement.

Joshua Smith, FounderRead the founding letter →
Orbit is our first model. More people world models are in development.About Halloway →

Watch Orbit read your company.

A 30-minute live briefing: how Orbit maps your teams, finds the people the market missed and answers the questions your leaders ask.

Book a briefing