Jerry Hsu
I am a passionate learner and always encourage myself to explore the possibilities of journey in my life.
Machine Learning Engineer
Taipei City | Taiwan
- National Taipei University | Master degree in statistics | 2017/09 - 2020/02
- National Sun Yat-sen University | Bachelor degree in mathematics | 2012/09 - 2016/06
Work Experience
CloudMile Inc. | Machine Learning Engineer | 2021/12 ~ Current | Taipei City
- Took charge in Hsinchu's Bigcity PLTV Precision Marketing project in the feature engineering and modeling
- Designed the retail related features and Machine Learning model.
- Developed the tree-based model in Python achieving 80% accuracy in classification task and 10% Decile-level MAPE in regression task.
- Collaborated with Data Engineer to complete the CDP platform by pandas and Looker studio.
- Co-lead with Solution Architect and Sales teams to deal with the pain points from overall clients
- Co-lead with Solution Architect and Sales in the project which was still in the business development stage and designed the Machine Learning solution in POC.
- Implemented the Machine Learning Solution in the POC (Proof of Concept) stage by Vertex AI
- Took the lead on the recommendation system’s internal project with 4 Machine Learning Engineers.
- Worked on the papers study and continued sharing among the ML team.
- Retrained GPT-2 model to replace the original model in the company’s product. (Still in the progress)
Academia Sinica CITI | Research Assistant | 2020/02 ~ 2021/12
- Reorganized and enhanced Automatic DJ system
- Proposed deep learning model surpassed the State of the Art model in the EDM subgenre classification task about 5% - 10% (Research Link)
- Refined the music structure annotation part to enable Automatic Disk Jockey System has ability to classify the segments of Electronic Dance Music into 6 different music structures.
- Converted the whole Automatic Disk Jockey System from Python 2 to Python 3.
- Contributed 2 AIDJ related researches in the AIDJ project (ICASSP, ISMIR)
- B.-Y. Chen, W.-H. Hsu, W.-H. Liao, M. A. M. Ram´ırez, Y. Mitsufuji, Y.-H. Yang, “Automatic DJ Transitions with Differentiable Audio Effects and Generative Adversarial Networks,” presented at the IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) (2022)
- A. Ramires et al., “The Freesound Loop Dataset and annotation tool,” in Proc. Int. Soc. Music Information Retrieval Conf., 2020
Skill & Certifications
Machine Learning skills
- Deep Learning (CNN, RNN, NN)
- Tree-based Model
- Music auto-tagging
- Predicted customer lifetime value
- Tabular data
Tools
- Python, R, SAS
- Numpy, Pandas, Scikit-learn
- GCP Vertex AI, AutoML
- Git, Airflow
Certrification
- Google Cloud Professional Machine Learning Engineer
- Google Cloud Professional Data Engineer
