Azure Storage

1.Designed and implemented multiple impactful Ads projects: Led the development of Happy Hour, Sponsored Alcohol, and SponsoredBrand projects, utilizing Kotlin, Redis, Elastic Search, Cassandra, and Kafka technologies. Increase Ads revenue around 15M in 2023.

2.Enhanced ad visibility: Designed and implemented Sponsored Brand feed and ad-exchange service integration, optimizing ad placements across various surfaces, including category and home pages. This led to an immediate revenue boost upon the M1 launch.

3.Implemented advanced sponsored brand ranking: Successfully executed complex sponsored brand ranking and auction logic using ML algorithms during the M2 launch.

4.Optimized system performance: Developed a sophisticated caching strategy leveraging Redis to minimize dependency calls, resulting in improved system reliability and scalability.

1.Design, Scope and Build scalable, robust inventory planning system to help Amazon optimize inventory balancing transfer using Java, Spring, AWS.

2.Cooperate with research scientist to implement and optimize multiple Linear Programming algorithms to help AMAZON optimize inventory transfer and help AMAZON saved 25 Million dollars in 2020.

3.Created and implemented Multiple-Lab-Architecture on Inventory planning team to run multiple A-B testing parallel simultaneously.

4.Scaling and Descaling on Inventory Planning team to adjust the traffic of AMAZON Prime Day.

• Implemented Python code and Apache Airflow DAG to periodically update 3 million cancer patients datasets.

• Built Healthcare resilient data pipelines in production and implemented the ETL/ELT to load a data warehousing via AWS

S3, Redshift.

• Implemented PostgreSQL query and utilized Flyway to automatically update schema and views in data warehouse.

• Improved data fetch server scalability 10 times via redesign REST API.

• Trained AWS machine learning models via SageMaker, and informed patients their sequencing DNA/RNA via SNS/SQS.

• Designed and owned US Disaster analytics and alert platform based on real-time twitter messages.

• Created automatic topic modeling data pipeline via NLTK, Spacy, Gensim library.

• Built disaster topic classification algorithm through random forest and RNN.

• Established Disaster Recommendation System based on retweet numbers and emotion detection obtained from IBM Waston

Tone Analyzer.