Product Data and Matching Engineer

About datawebot

datawebot gives AI agents ecommerce product data. Through one MCP server, an agent can search products across online shops worldwide, list each seller's price with a direct link to the shop, and run extraction Jobs on Amazon and Shopify stores. Anyone can sign in for free by approving an email. Learn more at https://www.datawebot.com

The role

The same pair of headphones appears under dozens of titles, sellers and variants across shops worldwide. You will build the matching and quality layer that groups listings into one product, separates new from used or resale offers, and flags results that do not fit the question.

What you will do

• Build product matching that groups listings across shops by brand, model, variant and identifiers such as GTIN or MPN.
• Use embeddings and LLMs where they help, with rules and evaluation sets that keep them honest.
• Classify offers by condition, seller type and country so agents can filter them.
• Create evaluation sets of real queries and measure match precision and recall over time.
• Turn quality findings into fixes in blueprints, search and offer ranking.

What you need

• Three or more years in data engineering, search, or applied machine learning on messy real-world data.
• Strong Python and SQL.
• Experience building evaluation sets and measuring quality, not only shipping models.
• Practical experience with LLM APIs or embeddings.
• Clear written English.

Nice to have

• Ecommerce catalog, product taxonomy or entity resolution experience.
• Experience with vector search or ranking.