Software engineer at Zoho, based in Madurai, India.

For the last four years I've worked on the search side of ManageEngine Log360 and EventLog Analyzer: storing, indexing and searching large volumes of security logs in Elasticsearch, often down at the Lucene level. Lately I've been putting LLMs on top of search.

• Moved Lucene FST term dictionaries off-heap, cutting JVM heap use by about 50%
• Switched high-cardinality filter fields to compressed bitmap indexes, cutting disk use by 40%
• Migrated log storage and search from Elasticsearch to ClickHouse for its OLAP capabilities, faster retrieval and better compression
• Wrote an Elasticsearch version facade that five other ManageEngine products now use
• Built a reindexing framework that moves live data across Elasticsearch versions without loss
• Designed sync and async REST search APIs, and AWS and GCP log collectors

• nlsearch: talk to Elasticsearch in plain English. An Elasticsearch plugin where an LLM turns a sentence into a real Elasticsearch call and explains the result. I wrote up how it works on my blog.

Search and data: Elasticsearch · ClickHouse · Lucene · PostgreSQL · MySQL · MSSQL · Redis · Kafka · Solr

Languages: Java · Python · JavaScript · SQL · Bash · Groovy · C

Backend and web: REST APIs · Tomcat · EmberJS · Django · HTML / CSS

Cloud and DevOps: Docker · Kubernetes · AWS (S3, CloudTrail) · GCP (Cloud Logging) · CI/CD · Git · Gradle · Linux

AI and ML: LLM tool use · RAG on Elasticsearch · LangChain4j · Ollama · OpenAI / Anthropic / Gemini APIs · TensorFlow

Testing: JUnit · integration testing · JMeter

Portfolio · Blog · LinkedIn · sheikmohammedsha@gmail.com

Open to agentic AI, LLM and search engineering roles, anywhere.