• Architecting the multi-agent reasoning engines that bridge foundation models with production-grade data applications. I lead the transition of GenAI into mission-critical systems, enabling users across Linkedin to conduct sophisticated, self-serve analytics through agentic workflows

• Data Governance & Democratization: Developing frameworks and standards that balance data accessibility with quality controls, enabling self-service analytics while reducing operational overhead

• Foundational Architecture: Designing scalable data architectures and migration strategies that support growing organizational needs and improve development velocity

• Cross-Functional Leadership: Driving technical initiatives across multiple teams and establishing engineering best practices

• Built Knowledge Graph to capture firm internal and external relationships to extract business insights; built data pipelines to combine multiple sources of structured and unstructured data into one heterogeneous graph with millions of vertices and billions of edges

• Built distributed and scalable solution in Hadoop ecosystem for serving graph data, reducing the response time of each batch API call by more than 80% on average - using pre-processing and caching the milestone data into sequence files

• Constructed framework in Hadoop ecosystem to allow indexing and generic filtering to enable querying the graph using attributes and secondary keys, helping reduce the number of calls by 50% compared to when only querying with primary keys was possible

• Led and implemented data governance standards for team's graph generation and analytics pipelines, resulting in streamlined processes and improved data integrity – helping comply with consent orders from different Govt. organizations

• Built an internal machine-learning platform for training models and improved speed of model training using Nvidia's RAPIDS.

• Developed ETL pipelines for data ingestion from 6 regional companies and trained ML models for pricing P&L insurance.

• Building ETL pipelines and workflows for internal and external data sources and

• Working with Fuzzy Matching module for data matching and Data Integration flows.

• Helping create Pattern Recognition and Machine Learning models

• Using Market basket analysis to make recommendations