Data Scientist III · F5 Inc
- Founding data scientist for F5's Global Services organization. Designed and implemented the core analytics framework for support operations across all channels, enabling data-driven decision-making and operational insights
- Designed, containerized, and deployed a ML based case search web app using Dash for front-end, open-source pre-trained bi-encoder model for embedding retrieval, Snowflake as the vector store, and Azure Container Apps for deployment
- Built a batch job for summarizing live chat transcripts using Hugging Face pipeline and Databricks workflow. Implemented topic modeling on the summarized output to identify top issues from support chats resulting in 30% reduction in volume
- Implemented key phrase extraction to tag support cases and identify areas for problem management, reducing case avoidance analysis time from weeks to days
- Migrated and refactored legacy Snowflake views into modular dbt models, improving code scalability and version control
- Developed denormalized tables that enabled 200+ users across the organization access support data for CX analytics
- Designed and analyzed A/B tests to evaluate the impact of Generative AI features on F5's support portal
- Built end-to-end data pipeline to ingest website survey data resulting in 60% faster resolution of customer feedback
- Upskilled 20+ members in the organization with SQL and Tableau to accelerate operational reporting capabilities