• Created an AI-powered ETL query expansion framework with Large Language Models (LLMs) to automate SQL query parsing, metadata extraction, and transformation optimization. Optimized ETL workflows by restructuring event data

• Designed and deployed Airflow DAGs to orchestrate data ingestion and transformation, streamlining data movement across Snowflake, AWS S3, and relational databases for scalability and fault tolerance

• Executed data quality checks in Data Build Tool (DBT), integrating custom tests, freshness monitoring, and CI/CD workflows

• Developed real-time monitoring dashboards in Snowflake and Sigma to monitor customer events for customer tracking events, enabling data-driven insights, operational monitoring and error tracking

• Developed expertise in fine-tuning and optimizing AI prompts, enhancing virtual assistant performance by improving language consistency, reducing inference latency, and minimizing hallucinations across diverse use cases.

• Designed and implemented the "MLLM as a Judge" evaluation framework, leveraging the Greedy Triplet Ranking (GTR) algorithm to automate the unbiased evaluation and ranking of large language models (LLMs) based on their performance.

• Spearheaded the transformation of an AI assistant from a single-product to a multi-product sales powerhouse by enhancing its performance through proactive customer needs analysis, a refined feature catalog, and tailored product recommendations.

• Automated end-to-end data pipelines for multi-product recommendation systems, including data generation, ingestion, and storage, significantly expanding the assistant's capabilities to recommend and switch between different product categories.

• Conducted in-depth comparative analyses of market-leading LLM products, translating findings into actionable insights and strategic recommendations that enhanced the virtual assistant’s conversational quality and user experience.

• Mastered Alibaba-powered data tools such as Brand Data Bank, Business Reference, effectively distilling valuable insights

• Applied deep analysis on consumer flows, employed sorting and filtering methods to customize and profile the crowd, fostering nuanced understanding of market dynamics and consumer behavior patterns

• Generated daily reports to trace the effect of brand crowd placement and extracted product features based on customer reviews

• Performed in-depth case studies on loyalty initiatives in Taobao's competitive marketplace, providing strategic advice to refine client brands' positioning and drive long-term growth