Built data infrastructure, migrating company from manual processes to automated data stack using BigQuery, DBT, and Airflow for daily batch and streaming pipelines. Designed and implemented connectors for 8+ data sources (Adjust, Meta Ads, Google Ads, Google Play, AppLovin, Shopify), centralizing marketing and revenue data into unified data warehouse. Created dashboards tracking revenue (Shopify), ad spend, and user acquisition metrics in Looker, enabling data-driven decision-making for leadership team. Managed team of 2 analysts, establishing data modeling standards and DBT best practices.

Developed Streamlit application replacing manual Looker dashboard creation, enabling clients to access customized reports without analyst intervention. Built production data pipelines using Apache Airflow, EC2, and PostgreSQL, automating report generation and email delivery, reducing manual workload by 60%. Designed and maintained dashboards tracking student engagement, revenue metrics, and operational KPIs for educational platform clients. Managed AWS data infrastructure (S3, EC2, PostgreSQL) supporting automated reporting workflows.

Built Streamlit application integrating LangChain and RAG systems to automatically generate ad variations, pause underperforming campaigns, and optimize Google Ads performance. Implemented vector search pipeline using FAISS and ChromaDB for semantic ad content generation, achieving 15% improvement in campaign conversion rates. Developed automated workflows connecting Google Ads API with AWS infrastructure for real-time campaign performance monitoring and optimization.

Performed customer journey analysis and funnel optimization using SQL, identifying performance trends and opportunities across multiple client campaigns. Supported statistical testing for ad creative and landing page variations, providing analysis and insights to guide optimization decisions. Developed executive dashboards using Looker Studio, delivering weekly performance reports and actionable insights for account management teams.