Building the ML backend for generative AI advertising product. Current responsibilities include:

• Experimentation for LLM model output quality.

• ML observability design and development.

• API design and development.

• Developed machine learning advertising service which combines multi-touch attribution and marketing mix modeling via transfer learning.

• Implementing multitenant model configuration and Snowflake to Databricks data ETL.

• Reduced advertising ROI prediction times by 80% by guiding migration to real-time deployment service.

• Instrumented key performance metrics to monitor the health of thousands Databricks jobs running daily.

Developed data ingestion pipelines and SQL driven custom reporting service.

• Created component responsible for applying row transformations and converting data to parquet format.

• Designed and implemented secrets storage for Kubernetes-hosted microservice.

• Streamlined developer and staging deployment with environment bootstrapping scripts.

• Lead maintenance and support for a product that manages millions of search ads from product sheets.

• Reduced service downtime by adding Slack-integrated monitoring using Prometheus and Grafana.

I created a bulk metadata editor integrated into a complex metadata schema to compile millions of image tags. This metadata editor leveraged React.js and optimized SQL queries to provide a high performance user interface and back-end. Afterwards, I implemented a video deduping algorithm prior to video crunching to save storage overhead.

Worked in a fast paced team of developers to enhance a web service that optimizes shipping logistics.

Improved business workflow by creating an intranet tool that provides administrators the ability to analyze freight shipping logistics.

Streamlined troubleshooting by extending the logging system to output session specific logs.