Hello, we are Yuki and Sho, machine learning engineers on the AI/LLM team at Mercari.In this tech blog, we dive into how

Mercari's platform is a resource for developers and technology enthusiasts, offering insights into engineering challenges, technology stack, and culture at Mercari. Through articles, tech talks, and open-source contributions, Mercari offers insights into building innovative and scalable technology solutions. Developers can learn about Mercari's engineering culture, agile methodologies, and contributions to the broader tech community.

Mercari Engineering

Yuki and Sho, machine learning engineers at Mercari, discuss fine-tuning the SigLIP Vision Language model for Mercari's product catalog to enhance the 'Visually Similar Items' feature. Through A/B testing, the fine-tuned model significantly improved tap rates and purchase counts. They detail the fine-tuning process, model deployment, and evaluation methods, highlighting the superior performance of the updated model in real-world applications.

Fine-tuned SigLIP Image Embeddings for Similar Looks Recommendation in a Japanese C2C Marketplace

Fine-tuning of the SigLIP model using product data