Author Image

Hi, I am Oscar

I |

Oscar CHEN

Software Engineer at FLYWHEEL Inc.

I am a passionate software engineer with 4 years of working experience. Currently aiming to grow as a Data Engineer and DevOps Engineer.

Experiences

1
Software Engineer
FLYWHEEL Inc.

Nov 2023 - Present, Tokyo, Japan

Growth By Data.

Responsibilities:
  • Accelerated progress in a chaotic project by implementing data transformation using dbt and Dagster. Simplified complex logic with visual representations to improve clarity and maintainability. Proactively initiated discussions with teammates to resolve confusion and drive project momentum.
  • Spearhead the design and implementation of deterministic encryption/decryption for client-sensitive data, carefully balancing methodologies to seamlessly integrate this feature into existing workflows and preparing transition guides for legacy data.
  • Develop and manage 200+ data pipeline jobs and sensors in Dagster, aligned with complex data dependencies to drive client success; contribute extensively to development logs and documentation to streamline team processes.
  • Design and build the ML pipeline for a PoC project, enabling frequent, automated experiments tailored to complex customer scenarios, utilizing tools like dbt, Dagster, and Bazel.
  • Enhance team productivity by optimizing CI processes throughout the organization, minimizing unexpected changes in Python lint, and significantly increasing linter speed in VS Code; simultaneously, refactor and remove 10 + unused configuration-related code in the monorepo to maintain code quality and maintainability
Software Engineer
Tenchijin Inc.

Oct 2020 - Oct 2023, Tokyo, Japan

We endeavor to optimize land use activity for the best of the human

Responsibilities:
  • Lead a 3-member team, architect and develop an internal data catalog, and enhance 30% data usage inside comoany, utilizing Golang, GraphQL and MongoDB. Collaborated with a frontend engineer to create a user-friendly UI. Mentor an intern on programming methodologies and best practices for API development.
  • Initiate, develop and maintain more than 20 ETL pipelines with Airflow and GDAL. Automate the ingestion of the latest satellite data into S3 datalake and transform it into formats fitting the need of the data science team.
  • Spearhead the creation of feature design document within the engineering team, enhancing company members’ understanding of new features and establishing efficient development practices.
  • Increase 30% of users by developing data models for backend system, and devising RESTful APIs for new features, such as Memo/Project on Compass Map leveraging Django, Postgres, Redis, Prometheus, Grafana, and the AWS stack (including EC2, S3, RDS, Lambda, and Route 53).
  • Proactively dockerize several projects inside company, optimizing the scalabity and deployment, saving 30% time of regular deployment with GitHub Action.
  • Develop a machine learning system for continuous model evaluation, and model deployment through API services. Collaborate closely with the data scientist, automating the duplicated procedure through integrating their preprocessing procedures and algorithms into the pipeline system.
  • Contribute to client’s success, providing valuable insights from satellite and weather data analysis on optimal areas for agriculture product or potential construction spots, and contributing to successful decision-making processes.
2

Education

University of Tokyo
2018-2020
Master in Spatial Information Science
Harbin Institute Of Technology
2014-2018
B.Sc. in Control Science And Engineering

Projects

Master Research
Author September 2018 - June 2020

Designedand implemented the preprocess procedure of trajectory data. Used machine learning models(e.g. LightGBM, XGBoost, LSTM etc.) to analyze the mobility pattern, and tried to find out the pattern when people will be in the stations. Compared the performance of traditional statistic methodologies with the modern machine learning methodologies on long-term time series predcition.

academic ML
2019 KDD Cup Competition Task1
Contributor April 2019 - June 2019

Participated in a big data competition focused on predicting user choices for Baidu Map navigation recommendations. I am responsible for cleaning and exploring large datasets, conducting feature engineering, and building a baseline model using LightGBM. Achieved a commendable rank of 143 out of 789 participants. The competition enhance my problem-solving ability in reality problems, and the enrich my knowledge in data analysis and modeling.

academic ML