Senior Product Engineer – Desktop & Microsoft Integrations

About the Role

We are building professional-grade desktop software that embeds AI directly into the tools knowledge workers already use every day. Our product runs locally, handles sensitive content, and lives alongside Word, Excel, Outlook, and the broader Microsoft 365 environment — so the experience has to feel native, trustworthy, and fast on Windows. This role is for an engineer who takes the Microsoft ecosystem seriously: someone who understands that ships on Windows is a different bar than runs on Windows. You will own the desktop client and its integrations end-to-end — architecture, packaging, signing, distribution, Office add-in surface, and the UX that ties advanced AI capabilities into workflows users already trust. You will work directly with our AI engineers and our agent-based development swarm. Comfort collaborating with AI tooling — not just consuming it — is part of the job.

What You'll Do

Own the desktop client. Architecture, performance, stability, installer experience, and update path for our cross-platform desktop application — with Windows as the priority surface.

Build and maintain Microsoft 365 integrations. Office Add-ins (Word, Excel, Outlook) using the Office JavaScript API and the manifest model, including AppSource publishing and the validation cycle that goes with it.

Make Windows feel first-class. Code signing, SmartScreen reputation, MSIX/installer behavior, Windows Defender interactions, auto-update reliability, multi-user installs, and the long tail of enterprise-Windows quirks. -

Design UX for long-running and AI-assisted workflows. Streaming responses, partial results, recoverable failures, and progress for operations that take seconds-to-minutes — not milliseconds.

Implement secure local storage and data-handling patterns. Local model artifacts, user content, credentials, and telemetry — with privacy and offline use as defaults, not afterthoughts.

Harden installation, updates, and recovery. First-run experience, version coherence between desktop and add-ins, graceful handling of partial-update states, diagnostics that a support engineer can actually use.

Collaborate with AI engineers and the agent swarm. Surface model and pipeline capabilities cleanly in the UI; work alongside AI-driven development tooling on a daily basis.

What We're Looking For

Required

5+ years building desktop or rich-client applications that real users depend on — not just web apps wrapped in a shell.

Deep Windows expertise. You've shipped signed, auto-updating Windows software to non-technical users and dealt with the consequences. You know what SmartScreen does on day one of a new certificate.

Strong Microsoft 365 / Office Add-in experience. Office.js, manifest model, taskpane and command surfaces, AppSource submission, and the realities of the Office runtime sandbox.

Proficiency with TypeScript/JavaScript and modern frameworks. React or equivalent on the UI side; comfort with Node-side desktop runtimes.

Experience integrating complex backend or AI-powered features into UX — streaming, cancellation, retry, and degraded-mode behavior.

Product mindset. You ship things that feel finished. You notice the half-second jank and the off-by-one in a tooltip.

Comfort in a fast-moving, early-stage environment where the spec is partly in code, partly in Slack, and partly in your head.

Strongly Preferred

Electron (or Tauri / similar) at production scale — including native modules, IPC patterns, and the security model.

Windows packaging and distribution: MSIX, MSI, code signing (EV and standard), auto-update frameworks, enterprise deployment via Intune/SCCM.

Office Add-in publishing pipeline: Partner Center, AppSource validation, manifest-as-contract patterns, sideloading for development

Document-centric applications — anything where the user's primary artifact is a file they care about.

Security- or privacy-sensitive products — products that ship to regulated industries, handle PII, or run in offline/air-gapped contexts.

Working alongside AI coding agents as a daily collaborator (Claude Code, Cursor, Copilot, or similar) — including reviewing agent-generated changes and shaping prompts/context for them.

Nice to Have

C#/.NET familiarity for native Windows surfaces or COM interop scenarios.

PowerShell for installation, deployment, and diagnostics scripting.

NativeWindows APIs (Win32, WinRT) where the cross-platform abstraction isn't enough.

Local LLM runtimes (llama.cpp, ONNX Runtime, DirectML) and the GPU/driver landscape on Windows.

Outlook add-in specifics — the trickiest of the Office Surfaces

How We Work

Small team. Senior bar. Short feedback loops. AI agents are part of the development environment, not a curiosity. Expect to read, review, and direct agent-generated code as part of normal work. We optimize for shipping software users trust over shipping fast. It works on my Mac is not a release criterion. Decisions are written down. We push back on ambiguity early.

Senior Product Engineer – Desktop Applications & Microsoft 365 Integration experience