Research Scientist

Arcade is building an AI physical product creation platform that uses natural language and generative AI to enable the design, purchase, and sale of custom manufacturable products. The Research Scientist will lead pioneering Vision-Language Model research, inventing architectures and training methods to evaluate product images for attributes such as captions, market value, and aesthetic quality. The role also includes dataset strategy, custom evaluation metrics, and collaboration with engineering teams to deploy research in production.

Responsibilities:

• Innovation & Invention: Pioneer novel neural architectures, loss functions, and multimodal integration techniques. We expect you to invent new AI technologies and methodologies to solve unprecedented challenges in visual product perception
• VLM Fine-Tuning & Adaptation: Lead the deep fine-tuning, adaptation, and structural optimization of state-of-the-art Vision-Language Models (e.g., Qwen-VLM, PaliGemma2 etc.) for targeted, highly complex computer vision tasks
• Complex Attribute Estimation: Develop the mathematical and architectural foundations to extract highly nuanced signals from product images, translating subjective concepts (like aesthetics and market value) into rigorous, learnable objectives
• Dataset Strategy: Guide the creation, curation, and algorithmic augmentation of specialized multimodal datasets required to teach foundation models novel, domain-specific concepts
• Evaluation & Metrics: Invent rigorous evaluation frameworks and custom metrics to accurately measure model performance on abstract tasks where standard academic benchmarks do not exist
• Cross-Functional Collaboration: Work closely with engineering teams to ensure your proprietary research and new technologies translate seamlessly into scalable, production-ready systems