I’m not looking.
Amble is.

The job market is built for people looking for jobs.Amble is built for people who aren’t.

It gives you a Rep: an agent that knows your work, listens to the market for you, and only speaks up when something deserves your time.

See why

Recruiting has worked the same way for decades.

A company publishes a job. People find it. They apply. Someone filters the applications.

The internet made it possible to apply to more jobs.
LinkedIn made everyone searchable.
ATSs made it possible to filter more people.
Then AI made applying almost free.

Now we can generate applications faster than anyone can meaningfully read them. Companies respond with more automation, more screening and more filters.

More applications. More filters. More noise.

applicationapplicationapplicationapplicationapplicationapplicationapplicationapplicationapplicationapplicationapplicationapplicationapplicationapplicationapplicationapplicationapplicationapplicationapplicationapplicationapplicationapplicationapplicationapplicationfilterapplicationapplicationfilterapplicationapplicationfilterapplicationapplicationfilterapplicationapplicationfilterapplicationapplicationfilterapplicationapplicationfilterapplicationapplicationfilterapplicationapplicationfilterapplicationapplicationfilterapplicationapplicationfilterapplicationapplicationfilterapplicationapplication

A better application keeps the same loop.

The application is the problem.

Only a few people are looking.

But you’re not. You’re working.

But some things would make you listen.

Not looking doesn’t mean not interested.

That’s what your Rep is for.

Aiko Tanaka

Product designer

Not looking

Daniel Okafor

Backend engineer

Open to work

Lucía Serrano

Data scientist

Not looking

Tom Weber

Staff engineer

Not looking

You

Busy building

Not looking

This week

Shipping the new release
Two 1:1s
A talk on Thursday

Would make you listen

the right problem

Would make you listen

more ownership

Would make you listen

remote

Would make you listen

the right number

Your Rep

Listening for

the right problem
more ownership
remote
the right number

Listening

Give everyone a Rep.

Your Rep knows what you’ve actually done, what you can demonstrate and what would make you move.

Then you go back to work.
Your Rep listens.
Most weeks, nothing happens.
That’s a feature.

When something deserves your attention, it comes to you, already filtered against your work, your conditions and the things you’ve taught your Rep to care about. It arrives with one sentence:

I found something worth your attention.

37 read this week

Senior ML EngineerAdtech✕ Office 5 days
Founding AI EngineerSeed✕ Solo on-call
Evaluation LeadKestrel✓ Worth your time

Your Rep

Listening

You

At work

1 worth your time

The old system waits for you to apply.Your Rep listens for you.

But first, your Rep needs to know you.

You’re more than a PDF.

A CV tells your Rep where you’ve worked. Your work shows what you’ve actually done.

Amble reads your CV, GitHub, LinkedIn and the work you choose to share. Then it asks about what it couldn’t find.

How your Rep gets to know you.

01 · Reads

It reads before it asks.

CV, LinkedIn, GitHub, your writing.

02 · Claims

Every claim keeps its source.

What it can’t source, it doesn’t say.

Aa a claimwhere it was read

03 · Evidence

Evidence adds up.

One source is a hint. Three is a pattern.

04 · Capabilities

Strength you can check.

And it names what it doesn’t know yet.

CVCV_Maya_Lindqvist_2026.pdf

Maya Lindqvist

AI Engineer · Stockholm

Lantern Labs · 2024 —

Led the evaluation pipeline for support agents.

Rebuilt retrieval after it failed an audit.

Fjord Data · 2020 — 2024

Shipped two production search systems.

Mentored two engineers.

GitHubmlindqvist

tern · eval harness for LLM agents · ★ 1.2k

rag-audit · retrieval regression tests

dotfiles

WebsiteWriting

Your eval set is lying to you

Graders are models too

LinkedInProfile

Talk · Measuring agents, PyData Berlin 2025

412 connections

LLM evaluation

Strong evidence3 pieces of evidence

Led the evaluation pipeline for support agents.CV

tern · eval harness for LLM agents · ★ 1.2kGitHub⚠ co-produced

Your eval set is lying to youWebsite

Retrieval systems

Good evidence3 pieces of evidence

Rebuilt retrieval after it failed an audit.CV

Shipped two production search systems.CV

rag-audit · retrieval regression testsGitHub

Technical writing

Good evidence2 pieces of evidence

Graders are models tooWebsite

Talk · Measuring agents, PyData Berlin 2025LinkedIn

I don’t know yet

Leading a team

Limited evidence

Mentoring two engineers isn’t leading a team. I can’t tell yet.

Your Rep reads what’s already there

See what yours could already know.

Type your GitHub username. Your browser asks GitHub directly; nothing reaches us.

No GitHub? In the beta, your Rep starts with your CV.

Your professional map

Everything on it, your Rep can back up.

Maya’s professional map4 sources · 1 interview

What you’ve demonstrated

LLM evaluationStrong evidenceCVGitHubWebsite
Retrieval systemsStrong evidenceCVGitHub

What it’s less sure about

Production scaleSome evidenceCV
Technical writingSome evidenceWebsiteLinkedIn

What it still doesn’t know

Leading a teamNo evidence yetMentoring two engineers isn’t leading a team.

When you want to move, your coach helps you do it well: what to say, who to talk to, and what your career.md already says for you.

Amble works for one person: you.

It learns what matters

You probably don’t know exactly what would make you move.

Neither does your Rep. Yet. Your choices teach your Rep what actually matters.

Founding AI Engineer
Seed · 6 people
- €115K + 1.2%
- Remote
- Solo on-call
YesYes, if…DependsNo
Maya“I won’t carry a pager alone again.”
Staff AI Engineer
Series C · 400 people
- €150K
- Berlin office
- 5 days a week
YesYes, if…DependsNo
Maya“…I can stay in Stockholm.”
Eval Lead
Series A · 40 people
- €118K + 0.3%
- Remote
- Owns all evals
YesYes, if…DependsNo
Maya“That’s the job I’d build myself.”

What I learned

Won’t considerSolo on-call
Weighs mostStaying remote
Floor~€115K for the right work

A yes is data. A no is better data.

I’ll listen for these →

Then it listens

Silence is a feature.

Your Rep does the job hunting. You get your time back.

Listening

Nothing worth bothering you with.I’m still listening.

Listening for

AI evaluation
Staff or senior
Remote, Europe
€115K+
No solo on-call

Three of these Maya never said. She showed them, in her choices.

Next, during the beta

Then it tells you

Then your phone buzzes.

When something makes it through your Rep, it comes to you, on your phone.

37 read this week · 1 worth your time

Senior ML EngineerAdtech · Series D✕ Office 5 days a week
Founding AI EngineerSeed · 6 people✕ Solo on-call
LLM Platform EngineerFintech · 300 people✕ €95K, below the floor
AI EngineerAgency✕ Not evaluation work
Staff AI EngineerSeries C · Berlin✕ Relocation required
Evaluation LeadKestrel · Series A · 40 people

Whenever an offer states a salary, you see it, and your Rep uses it.

Amble09:41

I found something worth your attention.

Evaluation Lead

Kestrel · Series A · 40 people

Remote, Europe
€125K + 0.4%
Owns evals end to end

Why I’m showing you this

They need someone to own LLM evaluation. It’s your strongest evidence.Your map · ●●●●
Remote in Europe, and no solo on-call.Your choices
€125K is above the floor you showed me.Your choices

One thing to consider

You’d be their first eval hire. That means setting direction, and I haven’t seen you do that yet.

On WhatsApp or Telegram

On WhatsApp or Telegram.Only what’s worth your time.And why it made the cut.

Eventually, the search itself goes away.

Companies will have agents of their own.
They’ll describe the work, the evidence they need, the conditions and the range.
Their agent will find the people who could actually do it.
Your Rep will decide whether it’s worth your attention.

Each side reveals more as interest grows on both.

Coming in v2

Then it talks for you

Your agent talks to theirs.

It answers with evidence, keeps what’s private private, and only brings you what both sides want.

Kestrel’s agent

Maya’s Rep

Can she own evaluation for our agents?
Yes. She built and ran Lantern’s eval pipeline, and defended its design.
✓ CV · GitHub · Interview
What’s her salary floor?
Your range works. The number stays with her.
Private · € ••••
We’d like to talk.
So would she. I’ll set it up.

Both sides want the conversation. Maya decides.

Remove applications.

Where we are

We’re starting with your Rep.

Now
Meet your Rep
Understands your work and what matters
It reads your work, asks what it can’t know, builds your professional map and coaches you on what’s still missing.
Soon
Stop searching
Finds opportunities worth your attention
It starts looking for you. The opportunities worth your attention reach your phone.
Then
Remove applications
Your Rep talks to theirs
Company agents talk directly to personal Reps. Applications disappear.

Yours

Your Rep knows a lot about you.
You should own all of it.

maya/career.md

# Work
## Lantern Labs
Role: AI Engineer · 2024 —[cv]
# Capabilities
## LLM evaluation
Evidence: strong[cv] [gh] [interview]
tern · plugin system⚠ co-produced, undefended
Caching strategy⊘ retracted
# Intent
Remote, Europe · €115K+ · no solo on-call[your choices]

A profile you can carry.

Read itCorrect itExport itDelete it

Private by default.
Every claim has a source.
You decide what leaves your Rep.

Gone means gone.

Not looking doesn’t mean not listening.