Oct 2025 – Present · 8 mos

• Architected and Created Databricks ConnectorHT, a VS Code extension for remote cluster execution, leveraging AI-assisted development tools to rapidly build and deploy the internal open-source solution.
• Authored two technical patents: a fully filed AI-integrated Big Data architecture for massive-scale processing (Patent Pending), and an automated testing framework for Data Engineering pipelines (filing in process).
• Expanded DIANA’s capabilities by integrating specialized AI Insights Agents (Summarization, Clustering) to process massive datasets, transitioning the business from static, predefined dashboards to dynamic, real-time data discovery that automatically uncovers hidden correlations.
• Led the creation of DIANA (Data Intelligence and Analytics Natural Assistant), empowering non-technical stakeholders to execute natural language database queries and on-the-fly KPI transformations, effectively eliminating manual data engineering bottlenecks and slashing cycle times from weeks to hours.
• Led the end-to-end architecture and development of a standardized, scalable stream data solution for vehicle EEL (Executable and Error Logging )logs, processing 10M+ events daily and translating complex business logic into real-time deliverables for key stakeholders.
• Architected a high-throughput data extraction pipeline utilizing fully distributed Databricks clusters, effectively managing massive I/O loads to orchestrate bulk data writes to PostgreSQL with zero performance bottlenecking.

Oct 2024 – Oct 2025 · 1 yr

• Deployed automated defect triage code on Azure Kubernetes Service, optimizing resource allocation and reducing server load by 40%.
• Engineered a Big Data pipeline converting 2M+ daily DLT binary logs to human-readable format using Apache Parquet optimizations.
• Developed custom metric visualizations for DLT logs in OpenSearch/ELK stack, reducing mean-time-to-diagnose by 25%.
• Built end-to-end ETL framework integrating multiple data sources authentication, achieving 5x faster ingestion via Spark optimizations and delivering unified insights through curated data layers.

Feb 2024 – Sep 2024 · 7 mos

• Developed and implemented a defect triage model, automating the manual ticket tagging process to route issues directly to the appropriate developer.
• Analyzed large DLT log files, achieving a 99% reduction in log line volume through advanced filtering and aggregation techniques.
• Applied AI/ML techniques to intelligently tag tickets by analyzing parameters such as descriptions and DLT log lines, improving issue categorization accuracy.
• Reduced ticket reassignment ("hip hopping") by approximately 90%, leading to a significant decrease in estimated resolution time.
• Deployed an intelligent, description-based issue tracking system using Azure Functions for efficient issue management and routing.
• Automated the entire process of triggering Azure Functions for description summarization and DLT log line reduction/analysis using a CI/CD pipeline.