Founding Engineer

About Matrix

Matrix is a personal cloud computer for the AI-agent era. It gives coding agents and assistants a persistent place to run, with terminals, repos, applications, memory, permissions, and access from every screen.

• The product spans an AI kernel, gateway, PostgreSQL data layer, per-user Linux computers, a browser shell, native desktop and mobile clients, and the infrastructure that keeps all of it available.
• We use Matrix to build Matrix. Your own agents and development environment will run on the product you are improving.
• This is category-defining systems and product work: the primitives for software that is operated by people and agents together.

The role

We are looking for an exceptional product-minded systems engineer to join the founding team and own consequential parts of Matrix end to end.

• You will work across TypeScript services, Linux runtimes, PostgreSQL, realtime protocols, security boundaries, and polished product surfaces.
• You will help decide what the architecture should become, then write the code, tests, migrations, instrumentation, and operational paths that make it real.
• You will speak with users, watch them work, and turn the sharpest problems into durable product primitives rather than one-off patches.

What you'll own

• Agent kernel primitives: durable sessions, model routing, tool execution, context, memory, and safe delegation.
• Persistent developer runtime: terminals, PTYs, session recovery, repositories, previews, files, and long-running work.
• Cloud computer lifecycle: provisioning, identity, updates, health, backups, recovery, and owner-controlled data.
• Product experiences across the Canvas-first web shell, desktop, mobile, and CLI.
• Realtime coordination between people, agents, devices, and background work.
• PostgreSQL and Kysely data models with correct transactions, concurrency control, and tenant boundaries.
• Security boundaries for authentication, permissions, sandboxing, network access, secrets, and untrusted inputs.
• Reliability and observability that let a tiny team operate many always-on computers with confidence.
• The engineering culture: sharp specs, tests first, small reviewable changes, direct feedback, and high standards without ceremony.

A week might include

• Tracing why an agent session resumed without the tool context it needed, then fixing the contract rather than the symptom.
• Shipping a feature from PostgreSQL migration and Hono route through realtime events to the Canvas UI.
• Profiling terminal streaming and making a remote session feel local on a poor mobile connection.
• Threat-modeling a new agent capability and designing the sandbox or approval path before exposing it.
• Talking with a founder using six coding agents, then simplifying the workflow that creates the most friction.
• Reviewing an agent-authored pull request, strengthening its invariants, and teaching the system to avoid the same mistake next time.
• Responding to an operational issue, repairing the failure mode, and leaving behind a useful alert and runbook.

How we work

• Own outcomes. You carry important work from the first question through production behavior.
• Stay close to reality. Read the code, reproduce the failure, inspect the data, and talk to the user.
• Design for agents and people. Both need clear state, recoverable actions, and understandable permissions.
• Use AI as leverage. Agents help us explore, implement, test, review, and document, but judgment stays with the engineer.
• Build secure defaults. Auth, isolation, validation, resource limits, and failure behavior are part of the feature.
• Prefer durable primitives. Solve the class of problem when the extra work earns its keep.
• Move quickly with evidence. Tests, metrics, and direct observation let us be fast without guessing.
• Care about the surface. A reliable system that feels confusing is not finished.

What we're looking for

• A track record of shipping substantial products or systems with unusually high personal ownership.
• Excellent TypeScript and strong fundamentals in APIs, data modeling, concurrency, networking, and distributed systems.
• Comfort moving between backend infrastructure and user-facing product code.
• Strong PostgreSQL instincts, including transactions, locking, migrations, query behavior, and operational failure modes.
• Fluency in Linux and the curiosity to work below framework abstractions when the problem demands it.
• Security judgment: you naturally look for trust boundaries, escalation paths, resource exhaustion, and data leaks.
• Product taste and the ability to reduce a complicated workflow to a clear interaction.
• Clear written communication and the ability to make architectural tradeoffs legible.
• High agency, low ego, and the desire to build closely with users and founders.
• Significant overlap with European working hours.

Helpful, but not required

• Experience with AI agent runtimes, model APIs, tool protocols, or coding agents.
• Experience with terminals, PTYs, SSH, zellij or tmux, WebSockets, and realtime streaming.
• Experience operating multi-tenant or per-customer cloud infrastructure.
• Experience with Hono, Next.js, React, Electron, Expo, or native platform clients.
• Experience building sandboxes, permission systems, secure gateways, or developer infrastructure.
• Meaningful open-source work, an early-stage startup, or another environment where you owned the whole problem.

We care about slope, judgment, and evidence that you can make hard systems simple. We do not filter for a particular degree, pedigree, or previous title.

What success looks like

• Users trust Matrix with work that continues after they close their laptop.
• Agent and terminal sessions recover predictably across disconnects, devices, deploys, and failures.
• New capabilities ship with clear permissions, bounded resources, useful observability, and safe failure modes.
• Product loops get faster because the underlying primitives are getting simpler and stronger.
• You own a meaningful technical area and raise the quality and pace of the whole team.
• Customer problems move directly into shipped improvements without losing context in handoffs.

What this role offers

• Top-of-market cash compensation, calibrated to experience and location.
• Meaningful equity and real ownership in the company you are helping create.
• Direct collaboration with both founders and a large voice in technical and product direction.
• Founding-team scope: choose important problems, shape the architecture, and see the full company context.
• The best AI development tools and your own always-on Matrix computer.
• A small, ambitious team with no layers between an insight, a decision, and shipped work.
• Remote work with substantial overlap with European working hours.

This role suits someone who wants the responsibility, learning speed, and ownership of founding a company while staying close to the craft of building the product.