• Built a robust data pipeline to create a high quality dataset, facilitating seamless integration between two business lines at Zillow. Enabled data science partners to drive an estimated $30 million revenue enhancement.

• Developed a Python cookiecutter for self-serving Airflow DAG generation. Increased developers efficiency and facilitated the construction of over 20 DAGs for the team.

• Collaborated with principal architects to architect an enterprise dimension model, establishing best practices for its implementation.

• Built comprehensive dimension models to seamlessly onboard critical L1 metrics through a semantic layer

• Created a self-service tool facilitating seamless onboarding of raw JSON data into Parquet-based Hive tables. Widely adopted by numerous customers for efficient data lake integration

• Developed a scalable, standardized Scala-based pipeline structure for the team, segregating business logic from input/output. Reduced dev time, enhanced MR review, and eased unit test integration

• Built high quality data pipelines to create business critical data products by working with data science partners and cross functional teams

• Designed, built and maintained scalable, serverless, near real-time data warehouse; Ingested ~5M transactions daily; Reduced analytical query execution time by 10x

• Collaborated with security and compliance team to make the data secure and GDPR compliant

• Implemented lambda architecture for real-time dashboard to monitor ML model and model features

• Created Spark applications which reduced time of daily SQL queries from 3 hours to under an hour

• Deployed neo4j and implemented Pregel Algorithm to leverage Graph analysis for fraud prevention

• Dockerized Shiny app dashboard and deployed on K8s cluster to enable CI/CD pipeline

• Implemented interactive web app to generate high/low-risk rules to streamline Decision Scientists workload

• Built an ETL pipeline to load the data in PostgreSQL data warehouse for data analysis

UN Project:

• Scraped dark web, social media data anonymously and maintained ETL pipeline for its sentimental analysis

• Migrated Apache Solr to Elasticsearch – Designed distributed multi-node cluster to improve reliability, availability, and performance; indexed over half billion data records in Elasticsearch for fast retrieval

• Developed and deployed six rich interactive visualization dashboards to find actionable insights

• Integrated React.js into all the dashboard websites to ensure high quality and reusability of code

• Maintained ETL pipeline on the stream of Tweets

Lockheed Martin Project:

• Worked on analytical NLP project for real-time discovery of issues, their frames, and evolutionary topic tracking

• Wrote batch script to automate project pipeline; incorporated hashing to reduce time complexity of string matching

• Created visualization website to explore the content of large set of news articles to leverage quick reading