Hands-on production experience in applied computer vision using Python and PyTorch.
Proven experience with segmentation or pose estimation and a history of shipping pipelines, not just research.
Strong focus on data quality and evaluation, beyond model training.
Ability to thrive in a fast-paced startup environment, particularly in San Francisco (hybrid work).
Responsibilities
Build and maintain preprocessing pipelines for try-on, including segmentation, pose estimation, human parsing, and body mapping.
Validate and standardize inputs to ensure clean, consistent data for downstream models.
Benchmark and integrate new vision models, assessing their performance on evaluation datasets.
Conduct dataset quality checks and address any issues that arise.
Collaborate with ML scientists to transition research into daily operational pipelines.
Benefits
Real ownership in a small team where your work directly impacts production for brand partners.
Engagement with challenging and visible problems, such as achieving accurate try-on across diverse individuals and garments.
Full Job Description
About the Role
SPREEAI builds AI virtual try-on that fashion brands embed in their sites and apps. Every try-on starts with the same question: did we read the person and the garment correctly? This role owns the computer vision pipelines that answer it, from preprocessing through evaluation, and keeps them reliable in production.
What you'll do
Build and maintain preprocessing pipelines for try-on: segmentation, pose estimation, human parsing, and body mapping
Validate and standardize inputs so downstream models get clean, consistent data
Benchmark and integrate new vision models, and measure how they perform on our own evaluation datasets
Run dataset quality checks and fix the failures they surface
Work with ML scientists to move research into pipelines that run every day
Qualifications
Hands-on production experience in applied computer vision with Python and PyTorch
Experience with segmentation or pose estimation and a record of shipping pipelines, not only research
Comfort owning data quality and evaluation, not just model training
Comfort working in a fast-moving startup, in San Francisco (hybrid)
Why Join SPREEAI?
Real ownership: a small team where the pipelines you build run in production for brand partners
Hard, visible problems: accurate try-on across a wide range of people and garments
The pay range for this role is:
145,000 - 180,000 USD per year (San Francisco, CA)
