Dynamic Sale Price - JAVA, Kafka, RESTful API• Collaborated with product managers to translate business requirements into technical specifications, enabling users to lock in the price for products during the sale.• Designed and implemented search system changes to support token-based pricing for seamless integration.• Led the end-to-end development and deployment of the feature, enhancing revenue during the sales.

A/B Capability in Search System - MySQL, JAVA• Developed highly customisable A/B testing infrastructure for the search system, enabling real-time experimentation of new features, facilitating data-driven decisions for feature rollouts.Cloud Migration - Kubernetes, Google Cloud Platform• Led the seamless migration of both stateful and stateless search services to Google Cloud Platform (GCP), ensuring improved scalability to support future business growth and traffic demands.Elite Brand Search - JAVA, HBase, RESTful API• Collaborated with the project managers to gain a comprehensive understanding of the project's goals through the Product Requirements Document.• Designed and built a novel feature allowing users to refine searches exclusively for premium brands, elevating search and browse experience.• Integrated this feature with premium ads services, resulting in a remarkable revenue surge for the company. Semantic Matching - Solr, Zookeeper, RESTful API• Implemented semantic search using Solr's Dense-Vector data structure in Java, giving search results aligned with users’ actual query intent.• Engineered and developed robust pipelines for indexing 100 million documents daily, while optimizing Solr cores to facilitate the swift retrieval of highly relevant products under 100ms.• Attained significant business impact by eliminating null searches, resulting in elevation of CTR and revenue growth.

Flash POC: Semantic Matching - Solr, Gradle• Researched and developed a proof-of-concept search feature, leveraging Solr’s HNSW graphs to fetch 1000 semantically matching products from a big-data pool of products accurately under 100 ms.• Sharded 50 million documents into multiple cores, in leader-follower design managed by the zookeeper, effectively managing each system to cater for up to 1000 QPS per follower.Intelligent Product Insights - Python, DropWizard• Developed and integrated the product page enrichment features using NLP and ML models to extract top keywords and summarize user reviews.• Optimized user experience by enabling review comparisons and in-depth product analysis, increasing user engagement.

• Predicted global market sentiments utilizing labor statistics and commerce data to help client estimate the impacts on stock price changes.• Built and trained FinBERT and BERT models to analyze people sentiments (in batch and single words); Retrieved and operated on commerce and laborstatistics data (USA) using the APIs and filtered them per the recency of occurrence.• Extracted keywords from retrieved data by leveraging TF-IDF method and dynamically displayed top 100 keyword-sentiment relation using D3.js.

• Designed and innovated a tool to assess percentage conformance between the implementation and Architecture Design of system softwares.• Created APIs to retrieve folder structures and file data from GitHub and Bitbucket repositories, facilitating seamless access to source code.• Parsed C++ code to extract architectural information and compared it to XMI files derived from Enterprise Architecture's UML diagrams.

• Delivered an end-to-end cost model analysis to help the client set up a pigment manufacturing plant in China. Created final analysis report, decks and provided insights to the client.• Conducted various market research studies and created reports for clients covering topics such as "Global Paper Cup Market Report and Forecast 2019- 2024", "Maltodextrin Market Report for EMEA Region" and "Global Vegetable Oil Market Research".• Conducted a Search Engine Optimization (SEO) for a website that successfully enhanced the "Domain Authority" and the "Page Authority" which helped lift the website appearance on Google search engine results.