that make decisions, take action and run operations autonomously
All open roles
Applied AI Engineer
AI demos always work. You build the ones that survive Monday, with genuinely messy data and people relying on them to do their job.
Apply for this rolearound the way each business works
no lock-in, no endless licences
nalto deploys AI inside mid-sized companies. For real, in production.
We do it through AI workers: agents that take over a job from start to finish. They don't just communicate through voice and email: they make decisions, take action, and run operations autonomously across entire enterprise systems. Around them, we build custom software designed for the way each business actually works.
Mid-sized companies know AI could transform how they operate. What they lack is the team to make it happen, and most assume custom AI is reserved for large corporations. We prove them wrong: one process at a time, in production, measured in hours given back to their teams.
Nothing off the shelf, nothing stuck in a demo. And everything we build belongs to the client: no lock-in, no endless licences.
We operate in Spain and France, and we're building our presence in the UAE.
The role
What you build
- Agents and LLM chains wired into the company's real data, awkward edge cases included.
- Evaluations before delivery: what isn't measured doesn't ship.
- Guardrails, error recovery and logs a human can actually read when it breaks at 7pm.
What we're looking for
- You've put an LLM system in production and seen what breaks.
- You can say no to a use case that isn't reliable, and offer the one that is.
- Evaluation interests you as much as the prompt does.
What it isn't No proofs of concept. If it doesn't reach users, it doesn't count.
Why join us
Real production
Every project ends with an AI worker live in production. Then on to the next.
A direct line to decision-makers
You work with CEOs, not five layers of middle management.
Variety
New industries, new processes and new problems with every client, across three markets.
Best-tool budget
Latest models, your editor, your stack. We pay for what makes you fast.
Weekly demos
Every deployment is shown live to the team. You see everyone's work, and everyone sees yours.
How we work
Production or nothing
A demo isn't a deliverable. If it's not running in a real business, it's not done.
Extreme ownership
Your AI worker, your client, your call. If something breaks, even outside your lane, you fix it or flag it. Nobody here waits for a ticket.
Craft over volume
We'd rather ship one AI worker that runs flawlessly for years than five that need babysitting. Quality compounds, duct tape doesn't.
Talent density
Small team, no passengers. The best argument wins, whoever makes it.
Urgency with focus
Speed is a promise, not a vibe. We move fast on what matters and refuse to be busy with what doesn't.
How we hire
The process
Four steps, two weeks: a thirty-minute call with a partner; a technical conversation about something you've already built; one real, paid day of work on an anonymised client process; an answer within five working days. No HR round, no personality test, no whiteboard algorithms.
You get an answer
A partner reads it, not an automated filter. We answer everyone, including when the answer is no.
Sounds like you?
Let's talk.
A two-minute form. A link to your LinkedIn or GitHub is enough, no CV needed.
Apply for this role