{"data":{"projects":{"edges":[{"node":{"frontmatter":{"title":"Enterprise RAG Platform on AWS Bedrock","company":"Workhuman","tech":["AWS Bedrock","OpenSearch","Terraform","Python"],"github":"","external":"","diagram":"/diagrams/rag-platform.svg"},"html":"<p>Designed and deployed a Retrieval-Augmented Generation (RAG) platform on Amazon Bedrock and OpenSearch, backed by enterprise knowledge bases. Automated provisioning with reusable Terraform modules to enable scalable enterprise GenAI adoption.</p>"}},{"node":{"frontmatter":{"title":"Real-Time Streaming Platform","company":"Workhuman","tech":["Apache Flink","Flink CDC","Kafka","Apache Iceberg"],"github":"","external":"","diagram":"/diagrams/streaming.svg"},"html":"<p>Built an end-to-end real-time streaming PoC and MVP using Apache Flink, Flink CDC (Oracle &#x26; PostgreSQL), Kafka, and Apache Iceberg to validate an enterprise streaming strategy. Developed a reusable Flink framework — job configs, sources/sinks, transformation patterns — that standardized streaming development and accelerated onboarding of new use cases.</p>"}},{"node":{"frontmatter":{"title":"Generic Data Ingestion Framework (GDIF)","company":"Workhuman","tech":["Scala","Amazon EMR","MSSQL"],"github":"","external":"","diagram":"/diagrams/gdif.svg"},"html":"<p>Expanded a reusable data ingestion framework in Scala on Amazon EMR, adding MSSQL connector support and incremental loading. Reduced compute consumption and improved data freshness across enterprise sources.</p>"}},{"node":{"frontmatter":{"title":"Global Search Data Product","company":"Workhuman","tech":["OpenSearch","Python"],"github":"","external":"","diagram":"/diagrams/global-search.svg"},"html":"<p>Delivered a global search data product spanning 25+ years of historical data, with optimized partitioning and redesigned personalization jobs feeding OpenSearch indexes — improving query performance, search relevance, and cost efficiency.</p>"}},{"node":{"frontmatter":{"title":"Federated Query Platform","company":"Bazaar Technologies","tech":["Trino","Apache Pinot","Apache Ranger"],"github":"","external":"","diagram":"/diagrams/federated-query.svg"},"html":"<p>Designed a federated query architecture using Trino for unified access across heterogeneous data sources, and scaled a 100+ TB distributed data platform. Added Apache Pinot for sub-second real-time analytics on fast-moving event data, with data security and governance enforced through Apache Ranger.</p>"}},{"node":{"frontmatter":{"title":"SheetBots & Slackbots","company":"Bazaar Technologies","tech":["PySpark","Apache Airflow"],"github":"","external":"","diagram":"/diagrams/sheetbots.svg"},"html":"<p>Designed automated SheetBots and Slackbots in PySpark that push hourly data updates straight to business teams, orchestrated end-to-end with Apache Airflow. Also built dashboards and visualizations in Apache Superset so non-technical users could self-serve analysis.</p>"}}]}}}