SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Salary: USD 250,000 - 400,000 / annual
Vizcom is a Series B design platform used by 700,000+ designers at companies like Nike, GM, New Balance, and Hasbro for sketching, rendering, 3D work, and material design. The company has raised $52M and is building AI-powered design tools.
As a Research Engineer focused on Preference Data, you'll architect the system that transforms design session data into training-grade preference datasets. This is a core technical role bridging product instrumentation, data pipelines, and ML research—not a support function.
You'll own the full stack: designing what signals the product captures (working with product engineers), building pipelines from canvas to warehouse to training sets, establishing dataset versioning and lineage standards, managing privacy and contractual boundaries across enterprise agreements, and creating feedback collection mechanisms when needed. Every example must be traceable to its origin; every training result must be reproducible from a dataset fingerprint months later.
The role recognizes that design sessions are branching trees, not sequences—designers fork, backtrack, and abandon directions. Your job is to capture that judgment in the data structure itself, encoding ambiguity honestly (unpicked ≠ disliked; abandoned ≠ rejected).
In your first 90 days, you'll map the existing event surface and warehouse, ship one new signal end-to-end into researchers' hands, and establish the data standards for everything that follows. Your customers are researchers one desk away; you'll see impact within days when a dataset you built enables questions nobody could ask before.
The company values research-log culture (negative results celebrated), reproducibility over vibes, publication of learnings, and hiring through paid work trials on real problems—not LeetCode.
