Data Engineer at DuPont, Nov 2023 ~ Present

Hsinchu, Taiwan

- Designed and implemented end-to-end machine learning architectures to predict chemical experiment results, leveraging Databricks, Azure Cloud, and on-premises servers. 

- Refactored and optimized the existing data ETL pipeline for real-time processing, using Azure Data Factory, Logic App, and custom Python programs.

- Collaborated with cross-functional teams located worldwide, delivering projects and tasks for diverse business units across different regions.

Data Engineer at iRent, Jun 2022 ~ Nov 2023

Taipei, Taiwan

- Designed data ingestion pipelines from data sources like on-premise databases, Azure EventHub, and APIs, processing 25M raw records update daily, using Azure Data Factory, Azure Function, Python async programming, APIs.

- Introduced Azure Data Factory's parameterized data pipeline to prevent manual tasks, simplify the pipeline structure, and cut down daily ETL processing time by 75%.

- Refactored data transformation program using Databricks and PySpark, reducing processing time from 8+ hours to less than 30 seconds at most.

- Designed ML projects and introduced MLOps to improve machine learning workflow, using MLflow, Pyspark MLlib, Azure Devops, and Databricks CLI.

- Introduced and implemented Databricks Auto Loader as an streaming ETL solution, which cut down processing time by 50% and prevented overheads like manually tracking job checkpoints.

- Served as an in-house Python trainer, training 7-8 members in the data team.

AI Engineer at Atelli, Nov 2019 ~ Jun 2022

Taipei, Taiwan

- Developed AI digital ads optimization programs across multiple digital ad platforms, resulting in saving clients' ad spend by 20% and increasing ROAS by 200%.

- Implemented data ingestion from 3 mainstream ad platforms, processing millions of raw records daily.

- Designed data pipelines for ETL, ad bidding calculation, model inference, and applying adjustments to ad platforms, using a combination of crontab, APIs, MySQL, and AWS Sagemaker.

- Implemented a deep learning pipeline for ad audiences recommendation, which included feature engineering, model training, and predicting the suitable audiences among a variety of labeled groups with 80% accuracy, using AWS S3 and Sagemaker and custom Python scripts.

- Leveraged clustering algorithm to simulate human agent ad operation, resulting in 95% similarity with the real actions.

- Built a deep CNN for predicting click through rates of ad creatives with MAE 0.8%.

Digital Product Manager at SOCIAL GEAR, Dec 2017 ~ Mar 2019

- Managed Facebook API based software products (Chatbot, Ad Tech, Creative system) 

- Arranged Chatbot scenario, conversation flow, and UX for clients' business goals 

- Responsible for oversea partnership development and local business development 

- Facebook ad operation

Digital Marketing Specialist at Payeasy,  Jul 2015 ~ Jul 2016

Taipei, Taiwan

Digital marketing and CRM.

Marketing Specialist at I-MEI Food Company, Oct 2013 ~ Mar 2015

Taipei, Taiwan

Product marketing and management.