At Hindustan Times Media (Shine.com), I own the distributed Solr/SolrCloud search backbone indexing 50M+ candidate and job records. I modernized production collections from Solr 6.x/7.x to 9.10.1 with zero downtime and implemented vector search with hybrid BM25 and dense-embedding ranking.
I restored legacy ranking parity after the move from TF-IDF to BM25 and re-engineered a proprietary Lucene similarity library for Lucene 9. I also reduced the schema on a 48M-document collection from 217 fields to 108.
I manage production MongoDB and MySQL clusters, and built Python pipelines processing and indexing 4M+ daily records. At Phoenix Techno Vision, I administered database systems for a freight and parcel tracking platform and automated maintenance tasks with Python and Bash.

