This document summarizes Flipkart's approach to building a real-time search index for e-commerce. It discusses the need for real-time search to address issues like showing out of stock products. It describes Flipkart's traffic volume and catalog size. It then covers challenges like high update rates and using microservices. It evaluates Apache SolrCloud and describes Flipkart's approach using a near real-time forward index with an inverted index for filtering to enable real-time sorting, filtering and querying with latency comparable to columnar data. The system processes over 150 signals and 50k updates per second in production.