R-16 · CONTINUUM

LIVE

Founding Agent Platform Engineer, Continuum

The graph is an API before it is a UI. Agents plan, re-sequence and flag drift on it natively, and nobody transcribes a status into a ticket.

FULL-TIME · FOUNDING · REMOTE · OVERLAPPING HOURS

Apply for this role

OPENINGS

1

One. When it is filled, this page comes down.

BASE SALARY

£220,000 to £370,000

The band, published. The offer is the same whether or not you push for more.

EQUITY

0.50% to 1.15% of Continuum

Founding equity in the company you would actually work on.

01

The problem this role exists for

Most tools bolted an assistant onto a product designed for humans. Continuum was built the other way round: work is stored as a graph with real edges, which is a structure an agent can reason over directly rather than by scraping a board. The opportunity is obvious and still largely untaken.

You would own that surface: the API and tool definitions agents operate through, the permission and audit model that makes an autonomous mutation something a team will actually allow, the evaluation that says whether an agent's re-sequencing was any good, and the drift detection that flags a plan quietly diverging from reality. Agent engineering with a real substrate underneath it rather than a chat box on the side.

More on the company itself: Continuum

A board versus a dependency graphColumns of cards hide the relationships between work items. The same work drawn as a directed acyclic graph exposes the critical path and the blast radius of a delay.WHAT A BOARD SHOWS YOUPosition without structure.WHAT THE SAME WORK ACTUALLY ISCritical path in red. Delay one node and the rest is computed.
Figure 1: A board reports position; a graph reports consequence. The same work, drawn two ways. Only one of them tells you what happens if something slips, and almost nobody has heard the argument yet.

02

Your first week, and a normal Tuesday

Week one: point an agent at a real customer graph in a sandbox, let it try to re-sequence a plan, and write down every way it went wrong. That document becomes the roadmap, and it will be longer than anybody expects.

A normal Tuesday: tool and API design, an evaluation harness run against yesterday's changes, and a conversation about permissions, because the difference between a useful agent and an unusable one is almost always what it is allowed to do without asking. You will spend real time with customers who are nervous about this, and they are right to be.

WHAT YOU WOULD BE WORKING IN

TypeScript, the graph API underneath, whichever model providers are best that month reached through a gateway rather than pinned, and an evaluation harness you will largely build yourself. Deliberately unattached to any single model or framework.

We do not screen on tools. Nothing above is a requirement, and a stack list is a poor proxy for whether somebody can do the work. If you have gone deep in something else and would arrive with a view about why half of this is the wrong choice, that is an argument we would like to have.

03

What you would own

Not a ticket queue and not a slice of somebody else’s roadmap. Four things with your name against them in the decision log, from the first fortnight.

01

The agent surface

Tools and API through which an agent operates on a real graph.

02

Permissions and audit

What an agent may change without asking, and the record of what it did.

03

Evaluation

Whether a re-sequencing was actually good, measured rather than assumed.

04

Drift

A plan quietly diverging from reality, flagged before the status meeting.

04

What we are looking for

MINIMUM

What we would need to see

  • You have shipped something using large language models that real users depended on, and can describe how it failed
  • Strong engineering fundamentals. This is an API and systems role before it is a modelling one
  • You have built evaluation for a non-deterministic system rather than relying on impressions
  • Strong TypeScript, or an equivalent typed language

PREFERRED

What would move you up the pile

  • Tool use, function calling or agentic architectures in production rather than in a demo
  • You have designed a permission or audit model somebody's security team signed off
  • Familiarity with MCP, agent protocols, or the current state of that tooling
  • Healthy scepticism about agents, which we treat as a qualification rather than a problem

Nothing in the preferred column is a filter, and neither list is applied before a human reads you. No degree is required for anything on this page, and we have hired people with doctorates and people with no degree at all into the same rooms. If you meet most of the minimum and not all of it, write anyway and say in the first paragraph which part you do not meet.

05

Pay, and everything around it

£220,000 to £370,000 base, plus 0.50% to 1.15% of Continuum, reviewed upward annually without you having to ask. The band is published here rather than discovered three conversations in, and the offer is the same whether or not you negotiate, because the band is the band.

Everything else is what the whole group gets and is not negotiated per offer: a holiday floor rather than a ceiling, health and dental and vision for your family, sixteen to twenty weeks of parental leave for every parent, a funded setup, monthly paid group meetups somewhere in the world, one required funded trip abroad each year, a protected day a week on the research ventures, and no non-competes.

The full list, the salary band chart and the visa sponsorship map are on the careers page. We sponsor without qualification across seventeen jurisdictions and publish exactly which.

06

Apply for this role

Four steps: a conversation with somebody who would actually work with you, one real problem drawn from work we are genuinely doing, a working session on your solution, and a decision within a week of that session. The problem is paid if it runs long, and you keep the work either way.

Or write to careers@cnsolarlabs.com.

What you send here reaches us and no one else. We use it to reply to you, we do not add you to any list, and you can ask us to delete it at any time. See the privacy policy.