2 months ago
Remote, United States or San Francisco, CA, USAMid Level
H1B sponsor
Base Salary
$139k - $286k/yr
Responsibilities
- Prototype new architectures for Pinterest vision-language models and fine-tune open-source LLMs.
- Improve visual perception and tool-use capabilities of multimodal models.
- Develop evaluation benchmarks for vision-centric capabilities such as fashion-style recommendations.
- Collect visual training data for Pinterest Canvas, including data for RLHF and targeted fine-tuning.
- Read research papers, participate in technical discussions, and help define the company’s visual generative strategy.
- Publish and publicize work through conferences, paper submissions, and blog posts.
Requirements
- At least 2 years of industry experience in computer vision.
- M.S. or PhD in machine learning, computer science, or a related field.
- Experience with generative computer vision models, preferably including visual encoders and LLMs.
- Preferred qualifications include publications at top machine learning conferences.
- Preferred experience includes Cursor, Copilot, Codex, or similar AI coding assistants.
- Familiarity with LLM-powered tools for documentation search, experiment analysis, SQL and data exploration, and engineering workflow acceleration is a plus.
Benefits
- Remote within the country, with in-person office collaboration required 1–2 times per quarter.
- Eligible for equity.
- Not eligible for relocation assistance.
Tech Stack
Categories
About Pinterest
Pinterest builds a visual discovery and bookmarking platform where people search, save, and shop ideas across home, fashion, food, and more. The company monetizes primarily through advertising and shopping integrations, using visual search and recommendation systems to connect users and merchants. Founded in 2010 and headquartered in San Francisco, Pinterest is a public company on the NYSE (PINS) with around 482 million monthly users worldwide.