Recommendations
LLMs for Relevance: Automating High-Quality Product Relevance Labeling in Flipkart Search
Our short summary of Flipkart’s engineering write-up. The full story, with all the detail, is on their blog.
Read the full write-up on blog.flipkart.tech (opens in a new tab)Problem
Manual relevance labeling was slow, expensive, and noisy.
Solution
They trained an LLM to judge query‑product relevance, first with supervised fine‑tuning and then with Grouped Relative Policy Optimization to improve consistency.
Matches or beats human annotators at a fraction of the cost
Want the architecture, code and numbers? They’re in the original post.
Read the full write-up on blog.flipkart.tech (opens in a new tab)