ML Engineer II — Visual AI & Multimodal LLMs

Pinterest; United States; Hybrid; USD 139,000 - 286,000; Full time

Job summary

Pinterest is seeking a research engineer/scientist to advance vision-centric LLMs and multimodal AI capabilities. You’ll join the Pinterest Labs visual modeling team to build production-ready models that perceive visual details and understand user aesthetics, collaborating with engineers and researchers across the company.

The role emphasizes hands-on work with generative computer vision models, various visual encoders and LLMs, and developing evaluation benchmarks for vision-centric features

Qualifications

• Experience with generative computer vision models.
• Proficiency in evaluating and deploying ML models in production.
• Strong collaboration and publication skills are a plus.

Responsibilities

• Prototype architectures for Pinterest VLMs and finetune open-source LLMs with vision.
• Develop evaluation benchmarks for vision-centric capabilities.
• Read papers and discuss research, shaping our visual generative strategy.
• Assist data collection for visual training and RLHF enhancements.
• Publish work via conferences, blogs, and other channels.

Skills

Generative computer vision

LLMs / visual encoders

Research & deployment experience

Education

MS/PhD in ML/CS or related

Job description

Pinterest is seeking a research engineer/scientist to advance vision-centric LLMs and multimodal AI capabilities. You’ll join the Pinterest Labs visual modeling team to build production-ready models that perceive visual details and understand user aesthetics, collaborating with engineers and researchers across the company.

The role emphasizes hands-on work with generative computer vision models, various visual encoders and LLMs, and developing evaluation benchmarks for vision-centric features