Jan 2025 – Present · 1 yr 4 mos · Hybrid

Feb 2023 – Jan 2025 · 1 yr 11 mos · Hybrid

• Executed end-to-end machine learning workflows: problem definition, data collection, preparation, model development, evaluation, and deployment on cloud infrastructure, ensuring reproducibility and optimal performance.
• Bid Optimizer Enhancement for Real-Time Bidding (RTB):
• Developed a bid optimizer model to reduce CPM and maintain conversion rates, outperforming legacy ML and heuristic models.
• Conducted data analysis, created a label adjustment heuristic, and optimized feature selection, achieving ~20–30% reduction in bid prices and improved CPM in testing.
• Deployed the model using Docker and Google Cloud, enabling daily retraining, resulting in a 110% increase in client cost savings and a 12% reduction in cost per click.
• Feature Engineering for RTB Models:
• Extracted, sourced, and prepared data for machine-learning readiness across RTB workflows targeting bid optimization, cost per action, and conversion.
• Predicted user interest metrics, including category and ad-specific interest percentages, active/inactive status, and recent engagement.
• Integrated zip code-level demographic and economic data using U.S. Census resources to enrich models.
• Viant Chatbot Development:
• Led the creation of an internal chatbot leveraging Streamlit, OpenAI GPT-4 Turbo, and Pinecone to improve information access for 100+ employees.
• Designed a robust data pipeline, optimized performance through prompt engineering, and deployed the chatbot with Docker on Cloud Run, eliminating third-party dependencies.
• Other Notable Achievements:
• Presented a GenAI information session attended by 100+ colleagues.
• Secured 3rd place in an Innovation Event for an Automated Code Review plug-in using the OpenAI API.
• Won 1st place on Bug Bash Day by resolving critical query bugs and recommending cost-saving data warehouse practices.