Senior Product Engineer

Agency Notice: We are not currently working with recruiting agencies for this role. Please do not contact Vizcom employees regarding this position. Any resumes submitted without a prior agreement will be considered unsolicited.

Vizcom is an AI design tool for industrial and product designers: from sketch to render, 3D model and video in one place. Teams at GM, Dior, HOKA, EA, and Roche use it, along with a big base of individual designers and students. We're Series B with 300+ enterprise customers and a small team.

Right now Vizcom is where the work gets made: 2D, 3D and video. Next, it needs to be where the product gets decided, with the review, the feedback and the context all staying in Vizcom instead of scattering across Miro, Slack and trackers. That means a faster canvas, real collaboration, and an agent that works inside the canvas instead of sitting next to it.

You'll own big pieces of that, end to end. You'll scope it with design, build the frontend and the API it needs, ship it, watch how designers actually use it, and fix what's wrong. We want someone who cares about how a brush stroke feels as much as whether the code is clean.

You'll report to our engineering manager and work every day with product, design, AI and backend engineers.

• Workbench performance. Our biggest user complaint is that the app gets "sluggish," especially in big files and long sessions. You'll profile frame time, memory and texture decode paths, set baselines, and keep it fast as files grow to hundreds of elements.

• Pro-grade 2D tools. Think brush engine, layer masks, selections, transform and warp. These need to feel as good as Photoshop or Procreate, because that's what our users compare us to.

• Agent-first canvas. Build the UI for create → review → present, with an AI agent that can see and act on the whole workbench. Make AI workflows (Modify, Regions, Animate, 3D) feel like one tool instead of a pile of features.

• Projects and collaboration. Projects that sit above files, real-time multiplayer, comments and review, and roles and permissions that actually make sense for enterprise teams.

• 3D in the browser. Three.js/WebGL scenes as first-class canvas elements, used as inputs to generation and in review.

• Quality. Write E2E tests, add instrumentation, and own your features in production. Everyone here takes on-call.
  

• 5+ years shipping complex frontend apps in React and TypeScript, the kind with deep state and lots of interaction, not CRUD dashboards.

• Real experience with Canvas, WebGL/WebGPU, or a custom rendering engine. You've chased dropped frames with a profiler and won.

• Comfortable across the stack. You can design a GraphQL schema, write the Postgres migration, and ship the feature without waiting on anyone.

• Product taste. You'll push back on a spec, prototype three options, and know which one feels right.

• You use AI coding tools heavily and still own every line. Speed is cheap now; we're hiring for judgment and verification.

• You own outcomes, not tickets.

Probably not a fit if: you want a fully specced backlog, you've only built forms and dashboards, or you think performance is someone else's job.

• You've built creative tools: drawing, image editing, design, CAD or 3D.

• Experience with CRDTs or real-time multiplayer sync.

• Three.js, WASM, or GPU shader work.

• Experience building agent UIs or LLM tool-calling.

• You've worked on Electron or another desktop app.

• You sketch or design yourself, or you've worked closely with industrial designers.

TypeScript, React, Vite, WebGL/Three.js on the frontend. Node, GraphQL (PostGraphile), Postgres, Redis, and BullMQ on the backend. Kubernetes, Pulumi, Datadog, Statsig and GitHub Actions for infra and ops.

• Location: San Francisco, hybrid (3 days/week in office)

• Interviews: Recruiter Screen (30m) → coding (60m) → product + architecture deep-dive (60m) → half-day onsite & Lunch with the team → references

• To apply: send links (GitHub, portfolio, anything you've shipped) and a few lines on the hardest frontend perf problem you've solved.

• Recruiter screen (30m): mutual fit, role and impact.

• Coding Session (60 m)

  Onsite

• Lunch with the team

• Architectural deep-dive (60 m):

• HM Team fit (45m)

• Product and design collaboration (60 m) .

• Coding Session (60 m):

• CTO chat

• References and offer.

If you want your work to be used daily by world-class designers and you like owning problems end to end, let’s talk. Apply with links (GitHub, LinkedIn, notable projects) and a short note on something hard you shipped.