Provide critical guidance and feedback to shape the next generation of the Azure platform and all aspects of Microsoft’s customer experience.

- Lead AI and LLM initiatives at SearchStax- Defined the company's AI product and engineering strategy- Conceptualized and executed a diverse array of PoCs in Retrieval-Augmented Generation (RAG), search re-ranking, metadata enrichment and question-answering.- Architected the AI engineering framework at SearchStax, recruited and mentored a team of AI engineers.- Facilitated the integration of AI tools across various business functions, leading to increased operational efficiency in Product and Customer Support.

- Led teams supporting IT operations for non-profit. - Led development of virtual real-time meeting platform during the Covid pandemic, scaling to hundreds of concurrent users online.

Lucene, Solr and Elasticsearch consultant. Big data integration with open-source search products. Query understanding, search relevance tuning and performance tuning.Worked with clients like Atlassian, CareerBuilder, Indeed, Yardi etc.

Led, recruited and built the search team from the ground-up which included data scientists and search engineers. Wore multiple hats including product, engineering and executive/strategy. Owned search roadmap, search product strategy and cross-functional product management involving search. Worked closely with the CEO to chart company's product roadmap which was heavily search-centric. Directed the development of query understanding, relevance algorithms and search quality systems including:- Tensorflow query classifier - query and document normalization- semantic query parser- automated search quality regression testing (MRR/NDCG)- BERT-based embeddings for neural search- query NER- misspelling corrections- query log analysis and mining Initiated and championed Vetted's internal ecommerce Knowledge Graph which encompassed tens of thousands of product categories, brands and product attributes, from requirements gathering, stakeholder alignment to rollout.Led use of GPT-3 for various product, search and indexing pipeline needs including:- GPT-3 embeddings for search science- product-attribute matching- attribute-attribute matching (for synonym detection)- product attribute enrichment- product classification

- Developed and maintained large-scale data processing pipelines using Cascading on Amazon EMR, optimizing workflows for performance and scalability.- Implemented and fine-tuned Solr for advanced search capabilities, leading to improved search relevancy, speed, and user experience.- Designed and executed custom Solr plugins to extend functionality and meet specific business requirements, enhancing the search experience.- Worked on analytics and NLP solutions, enabling enhanced data retrieval, sentiment analysis, and user insights.- Collaborated with cross-functional teams to integrate search and analytics solutions into various applications, driving data-driven decision-making.

- Architected, built, and maintained the entire application stack: Single-handedly managed the full development lifecycle, from backend infrastructure to frontend web application.- Custom-built crawler: Developed a robust, high-performance web crawler that efficiently processes over 1 million realtor listings per day. Implemented advanced techniques for data extraction, ensuring accuracy and completeness.- Backend data extraction: Designed and implemented scalable data extraction pipelines to handle vast amounts of data, ensuring timely and accurate data ingestion and processing.- Search engine development: Built a powerful search engine using Lucene, tailored to meet the specific needs of the application. Implemented complex search algorithms to provide fast, relevant, and comprehensive search results for users.- Frontend development: Developed a user-friendly, responsive web application that allows users to easily search, filter, and interact with the realtor listings.- Maintenance and optimization: Continuously monitored and optimized the entire stack for performance, reliability, and scalability. Implemented best practices for code quality, testing, and deployment.

Wore multiple hats, including - Technical Lead and ScrumMaster of professional services software team - gathered requirements, performed design and facilitated project implementation using agile techniques. Hired and managed a team of 5 developers.Pre-sales - addressed clients' technical concerns during product presentations, produced technical whitepapers, wrote tender documents and designed solutions that integrated with clients' existing IT infrastructure.Build Bitch - managed build scripts, setup and maintained Ant, Anthill, Maven, CVS, SSH and Mailman on Linux.Technical Achievements include:Independently developed a Lucene-based search and indexing engine, features include support for distributed searching, pluggable file handling, indexes directories and URLs, indexes PDF, HTML, MS Office, ZIP, GZIP, TAR and XML files.Extended WebWork and Velocity as framework layer, using Hibernate for persistence. It reduced development time of projects by ~3 weeks for a 3 month long project.Wrote services framework based on Avalon Framework.Developed various reusable services, includinga unified notification service to send emails, ICQ, MSN Messenger, Yahoo! Messenger and Jabber messages via a single APIa conversion service to allow for pluggable file conversion (e.g. DOC -> HTML, PDF -> HTML, etc)a file repository service for storing large (>10 million) number of files, with pluggable strategies for distributing the filesan XML-based validation framework which also does form generationAdded WebDAV hooks to the product to interface with WebDAV clients, e.g. Windows Explorer