Founding Data/ML Engineer

Reports to: Steffen Sommer, CTO and co-founder

About Palette

For the last decade, knowledge work drifted away from its original promise. It was supposed to be about thinking, creating, deciding. Instead it became status meetings, reporting, alignment rituals, and busywork.

AI changes that. But the speed has created a new problem: the individual got faster than the organization. More and more work happens inside sessions between humans and agents, and the context behind that work disappears before anyone else in the company can see it.

Palette is the shared context layer for teams and AI. We ingest signals from the tools teams already use (Slack, Linear, Notion, GitHub, calendar, email), turn them into a living map of the organization, and serve that context to the agents people work with every day.

Palette Desktop is the app where teams put that context to work. Point it at a folder your team shares on Google Drive or Dropbox, and run agents on Anthropic, OpenAI, Mistral, or a local model inside it. More providers ship in regularly. It opens those agents up to the whole company, including non-engineers, and the team picks the model that fits. We shipped Desktop in May, and it's already where most of our own work happens.

We're 6 people in Copenhagen. Four co-founders, two founding engineers. We work with design partners across leading startups, scaleups, and household-name brands. We recently raised our pre‑seed.

The role

You’ll own the brain of Palette: the signal engine.

Today it captures events and stores them as signals — enough for the briefs and context pages we ship now. Next is a queryable graph of activities: everything that happens across a company's tools, on one timeline, answerable by who, what, when, and meaning, by both people and AI agents. You’ll drive that, end to end.

This is a builder role. You work AI‑native — use the tools, move fast, own the output. If the AI writes 80% of the code, great; just reason about it and ship.

This is a founding role. You’ll define how Palette turns raw signals into shared context.

Who you are

The ideal person sits at the intersection of four hats:

• Data engineer. You own the pipelines at real volume, and take data privacy and security seriously.
• ML engineer. Embeddings, clustering, retrieval, vector search, LLMs as system components, evals.
• Data scientist. You spot the use cases and drive what we build next on top of the data.
• Software engineer. Comfortable across the stack, comfortable in TypeScript / Postgres.

You won’t be equally deep in all four — nobody is. But you’re genuinely strong in at least one and eager to grow into the rest. Right skills, but more importantly the right mindset: strong product sense, agency, curiosity. You think in relationships and retrieval, and you care about cost and quality. Bonus points for RAG and graph‑shaped data.

The stack

• Desktop: Tauri v2, Rust, TypeScript / React / TanStack
• Frontend: Next.js (migrating to TanStack)
• Backend: Hono
• AI agents: Mastra, plus the coding‑agent harnesses we run on‑device
• Data & infra: PostgreSQL, Redis, Inngest, Nango, Railway
• Plus: Linear, GitHub Actions, WorkOS, Cloudflare, Sentry, Incident.io, Requesty, Swarmia, PostHog, Atlas, and more

We’re bullish on TypeScript, but we’re constantly evaluating our approach and open to being challenged.

How we work

• Copenhagen office. Mostly in person, but around two days from home every week.
• We aim to stay a small and lean team.
• We’re all doers and care about our craftsmanship.
• We’re constantly exploring what AI‑native means to us.
• We’re obsessed with making something people actually want to use, so everyone at the company talks to customers regularly.

What you get

• Top‑tier salary plus warrants
• Solid office in Copenhagen
• The gear you need
• Dental insurance
• Lunch, snacks, and drinks at the office

How to apply

Apply on The Hub.